Merge branch 'cpacker:main' into main

2023-10-24 11:40:30 +05:30
parent d62e9ce62b 39b545f869
commit a2ac46f1f9
18 changed files with 1170 additions and 171 deletions
--- a/.github/workflows/main.yml
+++ b/.github/workflows/main.yml
@@ -23,5 +23,6 @@ jobs:
        python -m pip install --upgrade pip
        pip install -r requirements.txt

-    - name: Run main.py
-      run: python main.py
+    - name: Run main.py with input
+      run: |
+        echo -e "\n\n\nn" | python main.py
--- a/.gitignore
+++ b/.gitignore
@@ -79,3 +79,6 @@ dmypy.json

 # Pyre type checker
 .pyre/
+
+# MemGPT config files
+configs/
--- a/README.md
+++ b/README.md
@@ -5,7 +5,9 @@
 <div align="center">

 <strong>Try out our MemGPT chatbot on <a href="https://discord.gg/9GEQrxmVyE">Discord</a>!</strong>
-
+ 
+ <strong>⭐ NEW: You can now run MemGPT with <a href="https://github.com/cpacker/MemGPT/discussions/67">local LLMs</a> and <a href="https://github.com/cpacker/MemGPT/discussions/65">AutoGen</a>! ⭐ </strong>
+ 
 [![Discord](https://img.shields.io/discord/1161736243340640419?label=Discord&logo=discord&logoColor=5865F2&style=flat-square&color=5865F2)](https://discord.gg/9GEQrxmVyE)
 [![arXiv 2310.08560](https://img.shields.io/badge/arXiv-2310.08560-B31B1B?logo=arxiv&style=flat-square)](https://arxiv.org/abs/2310.08560)

@@ -75,13 +77,6 @@ Install dependencies:
 pip install -r requirements.txt
 ```

-Extra step for Windows:
-
-```sh
-# only needed on Windows
-pip install pyreadline3
-```
-
 Add your OpenAI API key to your environment:

 ```sh
@@ -115,6 +110,12 @@ python main.py --use_azure_openai

 To create a new starter user or starter persona (that MemGPT gets initialized with), create a new `.txt` file in [/memgpt/humans/examples](/memgpt/humans/examples) or [/memgpt/personas/examples](/memgpt/personas/examples), then use the `--persona` or `--human` flag when running `main.py`. For example:

+```sh
+# assuming you created a new file /memgpt/humans/examples/me.txt
+python main.py
+# Select me.txt during configuration process
+```
+-- OR --
 ```sh
 # assuming you created a new file /memgpt/humans/examples/me.txt
 python main.py --human me.txt
@@ -123,6 +124,11 @@ python main.py --human me.txt
 ### GPT-3.5 support
 You can run MemGPT with GPT-3.5 as the LLM instead of GPT-4:
 ```sh
+python main.py
+# Select gpt-3.5 during configuration process
+```
+-- OR --
+```sh
 python main.py --model gpt-3.5-turbo
 ```

@@ -130,7 +136,19 @@ python main.py --model gpt-3.5-turbo

 Please report any bugs you encounter regarding MemGPT running on GPT-3.5 to  https://github.com/cpacker/MemGPT/issues/59.

+### Local LLM support
+You can run MemGPT with local LLMs too. See [instructions here](/memgpt/local_llm) and report any bugs/improvements here https://github.com/cpacker/MemGPT/discussions/67.
+
 ### `main.py` flags
+```text
+--first
+  allows you to send the first message in the chat (by default, MemGPT will send the first message)
+--debug
+  enables debugging output
+```
+
+<details>
+<summary>Configure via legacy flags</summary>

 ```text
 --model
@@ -139,10 +157,6 @@ Please report any bugs you encounter regarding MemGPT running on GPT-3.5 to  htt
  load a specific persona file
 --human
  load a specific human file
--first
-  allows you to send the first message in the chat (by default, MemGPT will send the first message)
--debug
-  enables debugging output
 --archival_storage_faiss_path=<ARCHIVAL_STORAGE_FAISS_PATH>
  load in document database (backed by FAISS index)
 --archival_storage_files="<ARCHIVAL_STORAGE_FILES_GLOB_PATTERN>"
@@ -152,6 +166,8 @@ Please report any bugs you encounter regarding MemGPT running on GPT-3.5 to  htt
 --archival_storage_sqldb=<SQLDB_PATH>
  load in SQL database
 ```
+</details>
+

 ### Interactive CLI commands

@@ -160,8 +176,6 @@ These are the commands for the CLI, **not the Discord bot**! The Discord bot has
 While using MemGPT via the CLI (not Discord!) you can run various commands:

 ```text
-//
-  enter multiline input mode (type // again when done)
 /exit
  exit the CLI
 /save
@@ -293,6 +307,6 @@ Datasets used in our [paper](https://arxiv.org/abs/2310.08560) can be downloaded
 - [x] Integration tests
 - [x] Integrate with AutoGen ([discussion](https://github.com/cpacker/MemGPT/discussions/65))
 - [x] Add official gpt-3.5-turbo support ([discussion](https://github.com/cpacker/MemGPT/discussions/66))
+- [x] CLI UI improvements ([issue](https://github.com/cpacker/MemGPT/issues/11))
+- [x] Add support for other LLM backends ([issue](https://github.com/cpacker/MemGPT/issues/18), [discussion](https://github.com/cpacker/MemGPT/discussions/67))
 - [ ] Release MemGPT family of open models (eg finetuned Mistral) ([discussion](https://github.com/cpacker/MemGPT/discussions/67))
- [ ] CLI UI improvements ([issue](https://github.com/cpacker/MemGPT/issues/11))
- [ ] Add support for other LLM backends ([issue](https://github.com/cpacker/MemGPT/issues/18))
--- a/config.py
+++ b/config.py
@@ -0,0 +1,307 @@
+import glob
+import json
+import os
+import textwrap
+
+import interface
+
+import questionary
+
+from colorama import Fore, Style
+
+from typing import List, Type
+
+import memgpt.utils as utils
+
+from memgpt.personas.personas import get_persona_text
+from memgpt.humans.humans import get_human_text
+
+model_choices = [
+    questionary.Choice("gpt-4"),
+    questionary.Choice(
+        "gpt-3.5-turbo (experimental! function-calling performance is not quite at the level of gpt-4 yet)",
+        value="gpt-3.5-turbo",
+    ),
+]
+
+
+class Config:
+    personas_dir = os.path.join("memgpt", "personas", "examples")
+    humans_dir = os.path.join("memgpt", "humans", "examples")
+    configs_dir = "configs"
+
+    def __init__(self):
+        self.load_type = None
+        self.archival_storage_files = None
+        self.compute_embeddings = False
+        self.agent_save_file = None
+        self.persistence_manager_save_file = None
+        self.host = os.getenv("OPENAI_API_BASE")
+        self.index = None
+
+    @classmethod
+    async def legacy_flags_init(
+        cls: Type["Config"],
+        model: str,
+        memgpt_persona: str,
+        human_persona: str,
+        load_type: str = None,
+        archival_storage_files: str = None,
+        archival_storage_index: str = None,
+        compute_embeddings: bool = False,
+    ):
+        self = cls()
+        self.model = model
+        self.memgpt_persona = memgpt_persona
+        self.human_persona = human_persona
+        self.load_type = load_type
+        self.archival_storage_files = archival_storage_files
+        self.archival_storage_index = archival_storage_index
+        self.compute_embeddings = compute_embeddings
+        recompute_embeddings = self.compute_embeddings
+        if self.archival_storage_index:
+            recompute_embeddings = False  # TODO Legacy support -- can't recompute embeddings on a path that's not specified.
+        if self.archival_storage_files:
+            await self.configure_archival_storage(recompute_embeddings)
+        return self
+
+    @classmethod
+    async def config_init(cls: Type["Config"], config_file: str = None):
+        self = cls()
+        self.config_file = config_file
+        if self.config_file is None:
+            cfg = Config.get_most_recent_config()
+            use_cfg = False
+            if cfg:
+                print(
+                    f"{Style.BRIGHT}{Fore.MAGENTA}⚙️ Found saved config file.{Style.RESET_ALL}"
+                )
+                use_cfg = await questionary.confirm(
+                    f"Use most recent config file '{cfg}'?"
+                ).ask_async()
+            if use_cfg:
+                self.config_file = cfg
+
+        if self.config_file:
+            self.load_config(self.config_file)
+            recompute_embeddings = False
+            if self.compute_embeddings:
+                if self.archival_storage_index:
+                    recompute_embeddings = await questionary.confirm(
+                        f"Would you like to recompute embeddings? Do this if your files have changed.\n    Files: {self.archival_storage_files}",
+                        default=False,
+                    ).ask_async()
+                else:
+                    recompute_embeddings = True
+            if self.load_type:
+                await self.configure_archival_storage(recompute_embeddings)
+                self.write_config()
+            return self
+
+        # print("No settings file found, configuring MemGPT...")
+        print(
+            f"{Style.BRIGHT}{Fore.MAGENTA}⚙️ No settings file found, configuring MemGPT...{Style.RESET_ALL}"
+        )
+
+        self.model = await questionary.select(
+            "Which model would you like to use?",
+            model_choices,
+            default=model_choices[0],
+        ).ask_async()
+
+        self.memgpt_persona = await questionary.select(
+            "Which persona would you like MemGPT to use?",
+            Config.get_memgpt_personas(),
+        ).ask_async()
+        print(self.memgpt_persona)
+
+        self.human_persona = await questionary.select(
+            "Which user would you like to use?",
+            Config.get_user_personas(),
+        ).ask_async()
+
+        self.archival_storage_index = None
+        self.preload_archival = await questionary.confirm(
+            "Would you like to preload anything into MemGPT's archival memory?"
+        ).ask_async()
+        if self.preload_archival:
+            self.load_type = await questionary.select(
+                "What would you like to load?",
+                choices=[
+                    questionary.Choice("A folder or file", value="folder"),
+                    questionary.Choice("A SQL database", value="sql"),
+                    questionary.Choice("A glob pattern", value="glob"),
+                ],
+            ).ask_async()
+            if self.load_type == "folder" or self.load_type == "sql":
+                archival_storage_path = await questionary.path(
+                    "Please enter the folder or file (tab for autocomplete):"
+                ).ask_async()
+                if os.path.isdir(archival_storage_path):
+                    self.archival_storage_files = os.path.join(
+                        archival_storage_path, "*"
+                    )
+                else:
+                    self.archival_storage_files = archival_storage_path
+            else:
+                self.archival_storage_files = await questionary.path(
+                    "Please enter the glob pattern (tab for autocomplete):"
+                ).ask_async()
+            self.compute_embeddings = await questionary.confirm(
+                "Would you like to compute embeddings over these files to enable embeddings search?"
+            ).ask_async()
+            await self.configure_archival_storage(self.compute_embeddings)
+
+        self.write_config()
+        return self
+
+    async def configure_archival_storage(self, recompute_embeddings):
+        if recompute_embeddings:
+            if self.host:
+                interface.warning_message(
+                    "⛔️ Embeddings on a non-OpenAI endpoint are not yet supported, falling back to substring matching search."
+                )
+            else:
+                self.archival_storage_index = (
+                    await utils.prepare_archival_index_from_files_compute_embeddings(
+                        self.archival_storage_files
+                    )
+                )
+        if self.compute_embeddings and self.archival_storage_index:
+            self.index, self.archival_database = utils.prepare_archival_index(
+                self.archival_storage_index
+            )
+        else:
+            self.archival_database = utils.prepare_archival_index_from_files(
+                self.archival_storage_files
+            )
+
+    def to_dict(self):
+        return {
+            "model": self.model,
+            "memgpt_persona": self.memgpt_persona,
+            "human_persona": self.human_persona,
+            "preload_archival": self.preload_archival,
+            "archival_storage_files": self.archival_storage_files,
+            "archival_storage_index": self.archival_storage_index,
+            "compute_embeddings": self.compute_embeddings,
+            "load_type": self.load_type,
+            "agent_save_file": self.agent_save_file,
+            "persistence_manager_save_file": self.persistence_manager_save_file,
+            "host": self.host,
+        }
+
+    def load_config(self, config_file):
+        with open(config_file, "rt") as f:
+            cfg = json.load(f)
+        self.model = cfg["model"]
+        self.memgpt_persona = cfg["memgpt_persona"]
+        self.human_persona = cfg["human_persona"]
+        self.preload_archival = cfg["preload_archival"]
+        self.archival_storage_files = cfg["archival_storage_files"]
+        self.archival_storage_index = cfg["archival_storage_index"]
+        self.compute_embeddings = cfg["compute_embeddings"]
+        self.load_type = cfg["load_type"]
+        self.agent_save_file = cfg["agent_save_file"]
+        self.persistence_manager_save_file = cfg["persistence_manager_save_file"]
+        self.host = cfg["host"]
+
+    def write_config(self, configs_dir=None):
+        if configs_dir is None:
+            configs_dir = Config.configs_dir
+        os.makedirs(configs_dir, exist_ok=True)
+        if self.config_file is None:
+            filename = os.path.join(
+                configs_dir, utils.get_local_time().replace(" ", "_").replace(":", "_")
+            )
+            self.config_file = f"{filename}.json"
+        with open(self.config_file, "wt") as f:
+            json.dump(self.to_dict(), f, indent=4)
+        print(
+            f"{Style.BRIGHT}{Fore.MAGENTA}⚙️ Saved config file to {self.config_file}.{Style.RESET_ALL}"
+        )
+
+    @staticmethod
+    def get_memgpt_personas(dir_path=None):
+        if dir_path is None:
+            dir_path = Config.personas_dir
+        all_personas = Config.get_personas(dir_path)
+        default_personas = [
+            "sam",
+            "sam_pov",
+            "memgpt_starter",
+            "memgpt_doc",
+            "sam_simple_pov_gpt35",
+        ]
+        custom_personas = list(set(all_personas) - set(default_personas))
+        return Config.get_persona_choices(
+            [p for p in custom_personas + default_personas], get_persona_text
+        ) + [
+            questionary.Separator(),
+            questionary.Choice(
+                f"📝 You can create your own personas by adding .txt files to {dir_path}.",
+                disabled=True,
+            ),
+        ]
+
+    @staticmethod
+    def get_user_personas(dir_path=None):
+        if dir_path is None:
+            dir_path = Config.humans_dir
+        all_personas = Config.get_personas(dir_path)
+        default_personas = ["basic", "cs_phd"]
+        custom_personas = list(set(all_personas) - set(default_personas))
+        return Config.get_persona_choices(
+            [p for p in custom_personas + default_personas], get_human_text
+        ) + [
+            questionary.Separator(),
+            questionary.Choice(
+                f"📝 You can create your own human profiles by adding .txt files to {dir_path}.",
+                disabled=True,
+            ),
+        ]
+
+    @staticmethod
+    def get_personas(dir_path) -> List[str]:
+        files = sorted(glob.glob(os.path.join(dir_path, "*.txt")))
+        stems = []
+        for f in files:
+            filename = os.path.basename(f)
+            stem, _ = os.path.splitext(filename)
+            stems.append(stem)
+        return stems
+
+    @staticmethod
+    def get_persona_choices(personas, text_getter):
+        return [
+            questionary.Choice(
+                title=[
+                    ("class:question", f"{p}"),
+                    ("class:text", f"\n{indent(text_getter(p))}"),
+                ],
+                value=p,
+            )
+            for p in personas
+        ]
+
+    @staticmethod
+    def get_most_recent_config(configs_dir=None):
+        if configs_dir is None:
+            configs_dir = Config.configs_dir
+        os.makedirs(configs_dir, exist_ok=True)
+        files = [
+            os.path.join(configs_dir, f)
+            for f in os.listdir(configs_dir)
+            if os.path.isfile(os.path.join(configs_dir, f))
+        ]
+        # Return the file with the most recent modification time
+        if len(files) == 0:
+            return None
+        return max(files, key=os.path.getmtime)
+
+
+def indent(text, num_lines=5):
+    lines = textwrap.fill(text, width=100).split("\n")
+    if len(lines) > num_lines:
+        lines = lines[: num_lines - 1] + ["... (truncated)", lines[-1]]
+    return "     " + "\n     ".join(lines)
--- a/interface.py
+++ b/interface.py
@@ -10,131 +10,172 @@ init(autoreset=True)
 # DEBUG = True  # puts full message outputs in the terminal
 DEBUG = False  # only dumps important messages in the terminal

+
 def important_message(msg):
-    print(f'{Fore.MAGENTA}{Style.BRIGHT}{msg}{Style.RESET_ALL}')
+    print(f"{Fore.MAGENTA}{Style.BRIGHT}{msg}{Style.RESET_ALL}")
+
+
+def warning_message(msg):
+    print(f"{Fore.RED}{Style.BRIGHT}{msg}{Style.RESET_ALL}")
+

 async def internal_monologue(msg):
    # ANSI escape code for italic is '\x1B[3m'
-    print(f'\x1B[3m{Fore.LIGHTBLACK_EX}💭 {msg}{Style.RESET_ALL}')
+    print(f"\x1B[3m{Fore.LIGHTBLACK_EX}💭 {msg}{Style.RESET_ALL}")
+

 async def assistant_message(msg):
-    print(f'{Fore.YELLOW}{Style.BRIGHT}🤖 {Fore.YELLOW}{msg}{Style.RESET_ALL}')
+    print(f"{Fore.YELLOW}{Style.BRIGHT}🤖 {Fore.YELLOW}{msg}{Style.RESET_ALL}")
+

 async def memory_message(msg):
-    print(f'{Fore.LIGHTMAGENTA_EX}{Style.BRIGHT}🧠 {Fore.LIGHTMAGENTA_EX}{msg}{Style.RESET_ALL}')
+    print(
+        f"{Fore.LIGHTMAGENTA_EX}{Style.BRIGHT}🧠 {Fore.LIGHTMAGENTA_EX}{msg}{Style.RESET_ALL}"
+    )
+

 async def system_message(msg):
-    printd(f'{Fore.MAGENTA}{Style.BRIGHT}🖥️ [system] {Fore.MAGENTA}{msg}{Style.RESET_ALL}')
+    printd(
+        f"{Fore.MAGENTA}{Style.BRIGHT}🖥️ [system] {Fore.MAGENTA}{msg}{Style.RESET_ALL}"
+    )
+

 async def user_message(msg, raw=False):
    if isinstance(msg, str):
        if raw:
-            printd(f'{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg}{Style.RESET_ALL}')
+            printd(f"{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg}{Style.RESET_ALL}")
            return
        else:
            try:
                msg_json = json.loads(msg)
            except:
                printd(f"Warning: failed to parse user message into json")
-                printd(f'{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg}{Style.RESET_ALL}')
+                printd(
+                    f"{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg}{Style.RESET_ALL}"
+                )
                return

-    if msg_json['type'] == 'user_message':
-        msg_json.pop('type')
-        printd(f'{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg_json}{Style.RESET_ALL}')
-    elif msg_json['type'] == 'heartbeat':
+    if msg_json["type"] == "user_message":
+        msg_json.pop("type")
+        printd(f"{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg_json}{Style.RESET_ALL}")
+    elif msg_json["type"] == "heartbeat":
        if DEBUG:
-            msg_json.pop('type')
-            printd(f'{Fore.GREEN}{Style.BRIGHT}💓 {Fore.GREEN}{msg_json}{Style.RESET_ALL}')
-    elif msg_json['type'] == 'system_message':
-        msg_json.pop('type')
-        printd(f'{Fore.GREEN}{Style.BRIGHT}🖥️ {Fore.GREEN}{msg_json}{Style.RESET_ALL}')
+            msg_json.pop("type")
+            printd(
+                f"{Fore.GREEN}{Style.BRIGHT}💓 {Fore.GREEN}{msg_json}{Style.RESET_ALL}"
+            )
+    elif msg_json["type"] == "system_message":
+        msg_json.pop("type")
+        printd(f"{Fore.GREEN}{Style.BRIGHT}🖥️ {Fore.GREEN}{msg_json}{Style.RESET_ALL}")
    else:
-        printd(f'{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg_json}{Style.RESET_ALL}')
+        printd(f"{Fore.GREEN}{Style.BRIGHT}🧑 {Fore.GREEN}{msg_json}{Style.RESET_ALL}")
+

 async def function_message(msg):
-
    if isinstance(msg, dict):
-        printd(f'{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}')
+        printd(f"{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}")
        return

-    if msg.startswith('Success: '):
-        printd(f'{Fore.RED}{Style.BRIGHT}⚡🟢 [function] {Fore.RED}{msg}{Style.RESET_ALL}')
-    elif msg.startswith('Error: '):
-        printd(f'{Fore.RED}{Style.BRIGHT}⚡🔴 [function] {Fore.RED}{msg}{Style.RESET_ALL}')
-    elif msg.startswith('Running '):
+    if msg.startswith("Success: "):
+        printd(
+            f"{Fore.RED}{Style.BRIGHT}⚡🟢 [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+        )
+    elif msg.startswith("Error: "):
+        printd(
+            f"{Fore.RED}{Style.BRIGHT}⚡🔴 [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+        )
+    elif msg.startswith("Running "):
        if DEBUG:
-            printd(f'{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}')
+            printd(
+                f"{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+            )
        else:
-            if 'memory' in msg:
-                match = re.search(r'Running (\w+)\((.*)\)', msg)
+            if "memory" in msg:
+                match = re.search(r"Running (\w+)\((.*)\)", msg)
                if match:
                    function_name = match.group(1)
                    function_args = match.group(2)
-                    print(f'{Fore.RED}{Style.BRIGHT}⚡🧠 [function] {Fore.RED}updating memory with {function_name}{Style.RESET_ALL}:')
+                    print(
+                        f"{Fore.RED}{Style.BRIGHT}⚡🧠 [function] {Fore.RED}updating memory with {function_name}{Style.RESET_ALL}:"
+                    )
                    try:
                        msg_dict = eval(function_args)
-                        if function_name == 'archival_memory_search':
-                            print(f'{Fore.RED}\tquery: {msg_dict["query"]}, page: {msg_dict["page"]}')
+                        if function_name == "archival_memory_search":
+                            print(
+                                f'{Fore.RED}\tquery: {msg_dict["query"]}, page: {msg_dict["page"]}'
+                            )
                        else:
-                            print(f'{Fore.RED}{Style.BRIGHT}\t{Fore.RED} {msg_dict["old_content"]}\n\t{Fore.GREEN}→ {msg_dict["new_content"]}')
+                            print(
+                                f'{Fore.RED}{Style.BRIGHT}\t{Fore.RED} {msg_dict["old_content"]}\n\t{Fore.GREEN}→ {msg_dict["new_content"]}'
+                            )
                    except Exception as e:
                        printd(e)
                        printd(msg_dict)
                        pass
                else:
                    printd(f"Warning: did not recognize function message")
-                    printd(f'{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}')
-            elif 'send_message' in msg:
+                    printd(
+                        f"{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+                    )
+            elif "send_message" in msg:
                # ignore in debug mode
                pass
            else:
-                printd(f'{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}')
+                printd(
+                    f"{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+                )
    else:
        try:
            msg_dict = json.loads(msg)
            if "status" in msg_dict and msg_dict["status"] == "OK":
-                printd(f'{Fore.GREEN}{Style.BRIGHT}⚡ [function] {Fore.GREEN}{msg}{Style.RESET_ALL}')
+                printd(
+                    f"{Fore.GREEN}{Style.BRIGHT}⚡ [function] {Fore.GREEN}{msg}{Style.RESET_ALL}"
+                )
        except Exception:
            printd(f"Warning: did not recognize function message {type(msg)} {msg}")
-            printd(f'{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}')
+            printd(
+                f"{Fore.RED}{Style.BRIGHT}⚡ [function] {Fore.RED}{msg}{Style.RESET_ALL}"
+            )
+

 async def print_messages(message_sequence):
    for msg in message_sequence:
-        role = msg['role']
-        content = msg['content']
+        role = msg["role"]
+        content = msg["content"]

-        if role == 'system':
+        if role == "system":
            await system_message(content)
-        elif role == 'assistant':
+        elif role == "assistant":
            # Differentiate between internal monologue, function calls, and messages
-            if msg.get('function_call'):
+            if msg.get("function_call"):
                if content is not None:
                    await internal_monologue(content)
-                await function_message(msg['function_call'])
+                await function_message(msg["function_call"])
                # assistant_message(content)
            else:
                await internal_monologue(content)
-        elif role == 'user':
+        elif role == "user":
            await user_message(content)
-        elif role == 'function':
+        elif role == "function":
            await function_message(content)
        else:
-            print(f'Unknown role: {content}')
+            print(f"Unknown role: {content}")
+

 async def print_messages_simple(message_sequence):
    for msg in message_sequence:
-        role = msg['role']
-        content = msg['content']
+        role = msg["role"]
+        content = msg["content"]

-        if role == 'system':
+        if role == "system":
            await system_message(content)
-        elif role == 'assistant':
+        elif role == "assistant":
            await assistant_message(content)
-        elif role == 'user':
+        elif role == "user":
            await user_message(content, raw=True)
        else:
-            print(f'Unknown role: {content}')
+            print(f"Unknown role: {content}")
+

 async def print_messages_raw(message_sequence):
    for msg in message_sequence:
--- a/main.py
+++ b/main.py
@@ -5,9 +5,11 @@ import glob
 import os
 import sys
 import pickle
-import readline
+
+import questionary

 from rich.console import Console
+
 console = Console()

 import interface  # for printing to terminal
@@ -18,56 +20,104 @@ import memgpt.presets as presets
 import memgpt.constants as constants
 import memgpt.personas.personas as personas
 import memgpt.humans.humans as humans
-from memgpt.persistence_manager import InMemoryStateManager, InMemoryStateManagerWithPreloadedArchivalMemory, InMemoryStateManagerWithFaiss
+from memgpt.persistence_manager import (
+    InMemoryStateManager,
+    InMemoryStateManagerWithPreloadedArchivalMemory,
+    InMemoryStateManagerWithFaiss,
+)
+
+from config import Config

 FLAGS = flags.FLAGS
 flags.DEFINE_string("persona", default=None, required=False, help="Specify persona")
 flags.DEFINE_string("human", default=None, required=False, help="Specify human")
-flags.DEFINE_string("model", default=constants.DEFAULT_MEMGPT_MODEL, required=False, help="Specify the LLM model")
-flags.DEFINE_boolean("first", default=False, required=False, help="Use -first to send the first message in the sequence")
-flags.DEFINE_boolean("debug", default=False, required=False, help="Use -debug to enable debugging output")
-flags.DEFINE_boolean("no_verify", default=False, required=False, help="Bypass message verification")
-flags.DEFINE_string("archival_storage_faiss_path", default="", required=False, help="Specify archival storage with FAISS index to load (a folder with a .index and .json describing documents to be loaded)")
-flags.DEFINE_string("archival_storage_files", default="", required=False, help="Specify files to pre-load into archival memory (glob pattern)")
-flags.DEFINE_string("archival_storage_files_compute_embeddings", default="", required=False, help="Specify files to pre-load into archival memory (glob pattern), and compute embeddings over them")
-flags.DEFINE_string("archival_storage_sqldb", default="", required=False, help="Specify SQL database to pre-load into archival memory")
+flags.DEFINE_string(
+    "model",
+    default=constants.DEFAULT_MEMGPT_MODEL,
+    required=False,
+    help="Specify the LLM model",
+)
+flags.DEFINE_boolean(
+    "first",
+    default=False,
+    required=False,
+    help="Use -first to send the first message in the sequence",
+)
+flags.DEFINE_boolean(
+    "debug", default=False, required=False, help="Use -debug to enable debugging output"
+)
+flags.DEFINE_boolean(
+    "no_verify", default=False, required=False, help="Bypass message verification"
+)
+flags.DEFINE_string(
+    "archival_storage_faiss_path",
+    default="",
+    required=False,
+    help="Specify archival storage with FAISS index to load (a folder with a .index and .json describing documents to be loaded)",
+)
+flags.DEFINE_string(
+    "archival_storage_files",
+    default="",
+    required=False,
+    help="Specify files to pre-load into archival memory (glob pattern)",
+)
+flags.DEFINE_string(
+    "archival_storage_files_compute_embeddings",
+    default="",
+    required=False,
+    help="Specify files to pre-load into archival memory (glob pattern), and compute embeddings over them",
+)
+flags.DEFINE_string(
+    "archival_storage_sqldb",
+    default="",
+    required=False,
+    help="Specify SQL database to pre-load into archival memory",
+)
 # Support for Azure OpenAI (see: https://github.com/openai/openai-python#microsoft-azure-endpoints)
-flags.DEFINE_boolean("use_azure_openai", default=False, required=False, help="Use Azure OpenAI (requires additional environment variables)")
+flags.DEFINE_boolean(
+    "use_azure_openai",
+    default=False,
+    required=False,
+    help="Use Azure OpenAI (requires additional environment variables)",
+)


 def clear_line():
-    if os.name == 'nt':  # for windows
+    if os.name == "nt":  # for windows
        console.print("\033[A\033[K", end="")
    else:  # for linux
        sys.stdout.write("\033[2K\033[G")
        sys.stdout.flush()


-def save(memgpt_agent):
-    filename = utils.get_local_time().replace(' ', '_').replace(':', '_')
+def save(memgpt_agent, cfg):
+    filename = utils.get_local_time().replace(" ", "_").replace(":", "_")
    filename = f"{filename}.json"
-    filename = os.path.join('saved_state', filename)
+    filename = os.path.join("saved_state", filename)
    try:
        if not os.path.exists("saved_state"):
            os.makedirs("saved_state")
        memgpt_agent.save_to_json_file(filename)
        print(f"Saved checkpoint to: {filename}")
+        cfg.agent_save_file = filename
    except Exception as e:
        print(f"Saving state to {filename} failed with: {e}")

    # save the persistence manager too
-    filename = filename.replace('.json', '.persistence.pickle')
+    filename = filename.replace(".json", ".persistence.pickle")
    try:
        memgpt_agent.persistence_manager.save(filename)
        print(f"Saved persistence manager to: {filename}")
+        cfg.persistence_manager_save_file = filename
    except Exception as e:
        print(f"Saving persistence manager to {filename} failed with: {e}")
+    cfg.write_config()


 def load(memgpt_agent, filename):
    if filename is not None:
-        if filename[-5:] != '.json':
-            filename += '.json'
+        if filename[-5:] != ".json":
+            filename += ".json"
        try:
            memgpt_agent.load_from_json_file_inplace(filename)
            print(f"Loaded checkpoint {filename}")
@@ -75,8 +125,12 @@ def load(memgpt_agent, filename):
            print(f"Loading {filename} failed with: {e}")
    else:
        # Load the latest file
-        print(f"/load warning: no checkpoint specified, loading most recent checkpoint instead")
-        json_files = glob.glob("saved_state/*.json")  # This will list all .json files in the current directory.
+        print(
+            f"/load warning: no checkpoint specified, loading most recent checkpoint instead"
+        )
+        json_files = glob.glob(
+            "saved_state/*.json"
+        )  # This will list all .json files in the current directory.

        # Check if there are any json files.
        if not json_files:
@@ -91,12 +145,16 @@ def load(memgpt_agent, filename):
                print(f"Loading {filename} failed with: {e}")

    # need to load persistence manager too
-    filename = filename.replace('.json', '.persistence.pickle')
+    filename = filename.replace(".json", ".persistence.pickle")
    try:
-        memgpt_agent.persistence_manager = InMemoryStateManager.load(filename)  # TODO(fixme):for different types of persistence managers that require different load/save methods
+        memgpt_agent.persistence_manager = InMemoryStateManager.load(
+            filename
+        )  # TODO(fixme):for different types of persistence managers that require different load/save methods
        print(f"Loaded persistence manager from {filename}")
    except Exception as e:
-        print(f"/load warning: loading persistence manager from {filename} failed with: {e}")
+        print(
+            f"/load warning: loading persistence manager from {filename} failed with: {e}"
+        )


 async def main():
@@ -104,93 +162,199 @@ async def main():
    logging.getLogger().setLevel(logging.CRITICAL)
    if FLAGS.debug:
        logging.getLogger().setLevel(logging.DEBUG)
-    print("Running... [exit by typing '/exit']")
+
+    if any(
+        (
+            FLAGS.persona,
+            FLAGS.human,
+            FLAGS.model != constants.DEFAULT_MEMGPT_MODEL,
+            FLAGS.archival_storage_faiss_path,
+            FLAGS.archival_storage_files,
+            FLAGS.archival_storage_files_compute_embeddings,
+            FLAGS.archival_storage_sqldb,
+        )
+    ):
+        interface.important_message("⚙️ Using legacy command line arguments.")
+        model = FLAGS.model
+        if model is None:
+            model = constants.DEFAULT_MEMGPT_MODEL
+        memgpt_persona = FLAGS.persona
+        if memgpt_persona is None:
+            memgpt_persona = (
+                personas.GPT35_DEFAULT if "gpt-3.5" in model else personas.DEFAULT
+            )
+        human_persona = FLAGS.human
+        if human_persona is None:
+            human_persona = humans.DEFAULT
+
+        if FLAGS.archival_storage_files:
+            cfg = await Config.legacy_flags_init(
+                model,
+                memgpt_persona,
+                human_persona,
+                load_type="folder",
+                archival_storage_files=FLAGS.archival_storage_files,
+                compute_embeddings=False,
+            )
+        elif FLAGS.archival_storage_faiss_path:
+            cfg = await Config.legacy_flags_init(
+                model,
+                memgpt_persona,
+                human_persona,
+                load_type="folder",
+                archival_storage_files=FLAGS.archival_storage_faiss_path,
+                archival_storage_index=FLAGS.archival_storage_faiss_path,
+                compute_embeddings=True,
+            )
+        elif FLAGS.archival_storage_files_compute_embeddings:
+            print(model)
+            print(memgpt_persona)
+            print(human_persona)
+            cfg = await Config.legacy_flags_init(
+                model,
+                memgpt_persona,
+                human_persona,
+                load_type="folder",
+                archival_storage_files=FLAGS.archival_storage_files_compute_embeddings,
+                compute_embeddings=True,
+            )
+        elif FLAGS.archival_storage_sqldb:
+            cfg = await Config.legacy_flags_init(
+                model,
+                memgpt_persona,
+                human_persona,
+                load_type="sql",
+                archival_storage_files=FLAGS.archival_storage_sqldb,
+                compute_embeddings=False,
+            )
+        else:
+            cfg = await Config.legacy_flags_init(
+                model,
+                memgpt_persona,
+                human_persona,
+            )
+    else:
+        cfg = await Config.config_init()
+
+    interface.important_message("Running... [exit by typing '/exit']")
+    if cfg.model != constants.DEFAULT_MEMGPT_MODEL:
+        interface.warning_message(
+            f"⛔️ Warning - you are running MemGPT with {cfg.model}, which is not officially supported (yet). Expect bugs!"
+        )

    # Azure OpenAI support
    if FLAGS.use_azure_openai:
-        azure_openai_key = os.getenv('AZURE_OPENAI_KEY')
-        azure_openai_endpoint = os.getenv('AZURE_OPENAI_ENDPOINT')
-        azure_openai_version = os.getenv('AZURE_OPENAI_VERSION')
-        azure_openai_deployment = os.getenv('AZURE_OPENAI_DEPLOYMENT')
-        if None in [azure_openai_key, azure_openai_endpoint, azure_openai_version, azure_openai_deployment]:
-            print(f"Error: missing Azure OpenAI environment variables. Please see README section on Azure.")
+        azure_openai_key = os.getenv("AZURE_OPENAI_KEY")
+        azure_openai_endpoint = os.getenv("AZURE_OPENAI_ENDPOINT")
+        azure_openai_version = os.getenv("AZURE_OPENAI_VERSION")
+        azure_openai_deployment = os.getenv("AZURE_OPENAI_DEPLOYMENT")
+        if None in [
+            azure_openai_key,
+            azure_openai_endpoint,
+            azure_openai_version,
+            azure_openai_deployment,
+        ]:
+            print(
+                f"Error: missing Azure OpenAI environment variables. Please see README section on Azure."
+            )
            return

        import openai
+
        openai.api_type = "azure"
        openai.api_key = azure_openai_key
        openai.api_base = azure_openai_endpoint
        openai.api_version = azure_openai_version
        # deployment gets passed into chatcompletion
    else:
-        azure_openai_deployment = os.getenv('AZURE_OPENAI_DEPLOYMENT')
+        azure_openai_deployment = os.getenv("AZURE_OPENAI_DEPLOYMENT")
        if azure_openai_deployment is not None:
-            print(f"Error: AZURE_OPENAI_DEPLOYMENT should not be set if --use_azure_openai is False")
+            print(
+                f"Error: AZURE_OPENAI_DEPLOYMENT should not be set if --use_azure_openai is False"
+            )
            return

-    if FLAGS.model != constants.DEFAULT_MEMGPT_MODEL:
-      interface.important_message(f"Warning - you are running MemGPT with {FLAGS.model}, which is not officially supported (yet). Expect bugs!")
-
-    if FLAGS.archival_storage_faiss_path:
-        index, archival_database = utils.prepare_archival_index(FLAGS.archival_storage_faiss_path)
-        persistence_manager = InMemoryStateManagerWithFaiss(index, archival_database)
-    elif FLAGS.archival_storage_files:
-        archival_database = utils.prepare_archival_index_from_files(FLAGS.archival_storage_files)
-        print(f"Preloaded {len(archival_database)} chunks into archival memory.")
-        persistence_manager = InMemoryStateManagerWithPreloadedArchivalMemory(archival_database)
-    elif FLAGS.archival_storage_files_compute_embeddings:
-        faiss_save_dir = await utils.prepare_archival_index_from_files_compute_embeddings(FLAGS.archival_storage_files_compute_embeddings)
-        interface.important_message(f"To avoid computing embeddings next time, replace --archival_storage_files_compute_embeddings={FLAGS.archival_storage_files_compute_embeddings} with\n\t --archival_storage_faiss_path={faiss_save_dir} (if your files haven't changed).")
-        index, archival_database = utils.prepare_archival_index(faiss_save_dir)
-        persistence_manager = InMemoryStateManagerWithFaiss(index, archival_database)
+    if cfg.index:
+        persistence_manager = InMemoryStateManagerWithFaiss(
+            cfg.index, cfg.archival_database
+        )
+    elif cfg.archival_storage_files:
+        print(f"Preloaded {len(cfg.archival_database)} chunks into archival memory.")
+        persistence_manager = InMemoryStateManagerWithPreloadedArchivalMemory(
+            cfg.archival_database
+        )
    else:
        persistence_manager = InMemoryStateManager()

-    # Moved defaults out of FLAGS so that we can dynamically select the default persona based on model
-    chosen_human = FLAGS.human if FLAGS.human is not None else humans.DEFAULT
-    chosen_persona = FLAGS.persona if FLAGS.persona is not None else (personas.GPT35_DEFAULT if 'gpt-3.5' in FLAGS.model else personas.DEFAULT)
+    if FLAGS.archival_storage_files_compute_embeddings:
+        interface.important_message(
+            f"(legacy) To avoid computing embeddings next time, replace --archival_storage_files_compute_embeddings={FLAGS.archival_storage_files_compute_embeddings} with\n\t --archival_storage_faiss_path={cfg.archival_storage_index} (if your files haven't changed)."
+        )

-    memgpt_agent = presets.use_preset(presets.DEFAULT, FLAGS.model, personas.get_persona_text(chosen_persona), humans.get_human_text(chosen_human), interface, persistence_manager)
+    # Moved defaults out of FLAGS so that we can dynamically select the default persona based on model
+    chosen_human = cfg.human_persona
+    chosen_persona = cfg.memgpt_persona
+
+    memgpt_agent = presets.use_preset(
+        presets.DEFAULT,
+        cfg.model,
+        personas.get_persona_text(chosen_persona),
+        humans.get_human_text(chosen_human),
+        interface,
+        persistence_manager,
+    )
    print_messages = interface.print_messages
    await print_messages(memgpt_agent.messages)

-
    counter = 0
    user_input = None
    skip_next_user_input = False
    user_message = None
    USER_GOES_FIRST = FLAGS.first

-    if FLAGS.archival_storage_sqldb:
-        if not os.path.exists(FLAGS.archival_storage_sqldb):
-            print(f"File {FLAGS.archival_storage_sqldb} does not exist")
+    if cfg.load_type == "sql":  # TODO: move this into config.py in a clean manner
+        if not os.path.exists(cfg.archival_storage_files):
+            print(f"File {cfg.archival_storage_files} does not exist")
            return
        # Ingest data from file into archival storage
        else:
            print(f"Database found! Loading database into archival memory")
-            data_list = utils.read_database_as_list(FLAGS.archival_storage_sqldb)
+            data_list = utils.read_database_as_list(cfg.archival_storage_files)
            user_message = f"Your archival memory has been loaded with a SQL database called {data_list[0]}, which contains schema {data_list[1]}. Remember to refer to this first while answering any user questions!"
            for row in data_list:
                await memgpt_agent.persistence_manager.archival_memory.insert(row)
            print(f"Database loaded into archival memory.")

+    if cfg.agent_save_file:
+        load_save_file = await questionary.confirm(
+            f"Load in saved agent '{cfg.agent_save_file}'?"
+        ).ask_async()
+        if load_save_file:
+            load(memgpt_agent, cfg.agent_save_file)
+
    # auto-exit for
    if "GITHUB_ACTIONS" in os.environ:
        return

    if not USER_GOES_FIRST:
-        console.input('[bold cyan]Hit enter to begin (will request first MemGPT message)[/bold cyan]')
+        console.input(
+            "[bold cyan]Hit enter to begin (will request first MemGPT message)[/bold cyan]"
+        )
        clear_line()
        print()

    while True:
        if not skip_next_user_input and (counter > 0 or USER_GOES_FIRST):
-
            # Ask for user input
-            user_input = console.input("[bold cyan]Enter your message:[/bold cyan] ")
+            # user_input = console.input("[bold cyan]Enter your message:[/bold cyan] ")
+            user_input = await questionary.text(
+                "Enter your message:",
+                multiline=True,
+                qmark=">",
+            ).ask_async()
            clear_line()

-            if user_input.startswith('!'):
+            if user_input.startswith("!"):
                print(f"Commands for CLI begin with '/' not '!'")
                continue

@@ -201,34 +365,21 @@ async def main():

            # Handle CLI commands
            # Commands to not get passed as input to MemGPT
-            if user_input.startswith('/'):
-
-                if user_input == "//":
-                    print("Entering multiline mode, type // when done")
-                    user_input_list = []
-                    while True:
-                        user_input = console.input("[bold cyan]>[/bold cyan] ")
-                        clear_line()
-                        if user_input == "//":
-                            break
-                        else:
-                            user_input_list.append(user_input)
-
-                    # pass multiline inputs to MemGPT
-                    user_message = system.package_user_message("\n".join(user_input_list))
-
-                elif user_input.lower() == "/exit":
+            if user_input.startswith("/"):
+                if user_input.lower() == "/exit":
                    # autosave
-                    save(memgpt_agent=memgpt_agent)
+                    save(memgpt_agent=memgpt_agent, cfg=cfg)
                    break

                elif user_input.lower() == "/savechat":
-                    filename = utils.get_local_time().replace(' ', '_').replace(':', '_')
+                    filename = (
+                        utils.get_local_time().replace(" ", "_").replace(":", "_")
+                    )
                    filename = f"{filename}.pkl"
                    try:
                        if not os.path.exists("saved_chats"):
                            os.makedirs("saved_chats")
-                        with open(os.path.join('saved_chats', filename), 'wb') as f:
+                        with open(os.path.join("saved_chats", filename), "wb") as f:
                            pickle.dump(memgpt_agent.messages, f)
                            print(f"Saved messages to: {filename}")
                    except Exception as e:
@@ -236,10 +387,12 @@ async def main():
                    continue

                elif user_input.lower() == "/save":
-                    save(memgpt_agent=memgpt_agent)
+                    save(memgpt_agent=memgpt_agent, cfg=cfg)
                    continue

-                elif user_input.lower() == "/load" or user_input.lower().startswith("/load "):
+                elif user_input.lower() == "/load" or user_input.lower().startswith(
+                    "/load "
+                ):
                    command = user_input.strip().split()
                    filename = command[1] if len(command) > 1 else None
                    load(memgpt_agent=memgpt_agent, filename=filename)
@@ -265,17 +418,23 @@ async def main():
                    continue

                elif user_input.lower() == "/model":
-                    if memgpt_agent.model == 'gpt-4':
-                        memgpt_agent.model = 'gpt-3.5-turbo'
-                    elif memgpt_agent.model == 'gpt-3.5-turbo':
-                        memgpt_agent.model = 'gpt-4'
+                    if memgpt_agent.model == "gpt-4":
+                        memgpt_agent.model = "gpt-3.5-turbo"
+                    elif memgpt_agent.model == "gpt-3.5-turbo":
+                        memgpt_agent.model = "gpt-4"
                    print(f"Updated model to:\n{str(memgpt_agent.model)}")
                    continue

-                elif user_input.lower() == "/pop" or user_input.lower().startswith("/pop "):
+                elif user_input.lower() == "/pop" or user_input.lower().startswith(
+                    "/pop "
+                ):
                    # Check if there's an additional argument that's an integer
                    command = user_input.strip().split()
-                    amount = int(command[1]) if len(command) > 1 and command[1].isdigit() else 2
+                    amount = (
+                        int(command[1])
+                        if len(command) > 1 and command[1].isdigit()
+                        else 2
+                    )
                    print(f"Popping last {amount} messages from stack")
                    for _ in range(min(amount, len(memgpt_agent.messages))):
                        memgpt_agent.messages.pop()
@@ -304,14 +463,23 @@ async def main():
        skip_next_user_input = False

        with console.status("[bold cyan]Thinking...") as status:
-            new_messages, heartbeat_request, function_failed, token_warning = await memgpt_agent.step(user_message, first_message=False, skip_verify=FLAGS.no_verify)
+            (
+                new_messages,
+                heartbeat_request,
+                function_failed,
+                token_warning,
+            ) = await memgpt_agent.step(
+                user_message, first_message=False, skip_verify=FLAGS.no_verify
+            )

            # Skip user inputs if there's a memory warning, function execution failed, or the agent asked for control
            if token_warning:
                user_message = system.get_token_limit_warning()
                skip_next_user_input = True
            elif function_failed:
-                user_message = system.get_heartbeat(constants.FUNC_FAILED_HEARTBEAT_MESSAGE)
+                user_message = system.get_heartbeat(
+                    constants.FUNC_FAILED_HEARTBEAT_MESSAGE
+                )
                skip_next_user_input = True
            elif heartbeat_request:
                user_message = system.get_heartbeat(constants.REQ_HEARTBEAT_MESSAGE)
@@ -322,10 +490,10 @@ async def main():
    print("Finished.")


-if __name__ ==  '__main__':
+if __name__ == "__main__":

    def run(argv):
        loop = asyncio.get_event_loop()
        loop.run_until_complete(main())

-    app.run(run)
+    app.run(run)
--- a/memgpt/autogen/memgpt_agent.py
+++ b/memgpt/autogen/memgpt_agent.py
@@ -80,7 +80,7 @@ class MemGPTAgent(ConversableAgent):
    def pretty_concat(messages):
        """AutoGen expects a single response, but MemGPT may take many steps.
        
-        To accomadate AutoGen, concatenate all of MemGPT's steps into one and return as a single message.
+        To accommodate AutoGen, concatenate all of MemGPT's steps into one and return as a single message.
        """
        ret = {
            'role': 'assistant',
--- a/memgpt/local_llm/README.md
+++ b/memgpt/local_llm/README.md
@@ -0,0 +1,103 @@
+⁉️ Need help configuring local LLMs with MemGPT? Ask for help on [our Discord](https://discord.gg/9GEQrxmVyE) or [post on the GitHub discussion](https://github.com/cpacker/MemGPT/discussions/67).
+
+👀 If you have a hosted ChatCompletion-compatible endpoint that works with function calling, you can simply set `OPENAI_API_BASE` (`export OPENAI_API_BASE=...`) to the IP+port of your endpoint. **As of 10/22/2023, most ChatCompletion endpoints do *NOT* support function calls, so if you want to play with MemGPT and open models, you probably need to follow the instructions below.**
+
+🙋 Our examples assume that you're using [oobabooga web UI](https://github.com/oobabooga/text-generation-webui#starting-the-web-ui) to put your LLMs behind a web server. If you need help setting this up, check the instructions [here](https://github.com/oobabooga/text-generation-webui#starting-the-web-ui). More LLM web server support to come soon (tell us what you use and we'll add it)!
+
+---
+
+# How to connect MemGPT to non-OpenAI LLMs
+
+**If you have an LLM that is function-call finetuned**:
+  - Implement a wrapper class for that model
+    - The wrapper class needs to implement two functions:
+      - One to go from ChatCompletion messages/functions schema to a prompt string
+      - And one to go from raw LLM outputs to a ChatCompletion response
+  - Put that model behind a server (e.g. using WebUI) and set `OPENAI_API_BASE`
+
+```python
+class LLMChatCompletionWrapper(ABC):
+
+    @abstractmethod
+    def chat_completion_to_prompt(self, messages, functions):
+        """Go from ChatCompletion to a single prompt string"""
+        pass
+
+    @abstractmethod
+    def output_to_chat_completion_response(self, raw_llm_output):
+        """Turn the LLM output string into a ChatCompletion response"""
+        pass
+```
+
+## Example with [Airoboros](https://huggingface.co/jondurbin/airoboros-l2-70b-2.1) (llama2 finetune)
+
+To help you get started, we've implemented an example wrapper class for a popular llama2 model **finetuned on function calling** (Airoboros). We want MemGPT to run well on open models as much as you do, so we'll be actively updating this page with more examples. Additionally, we welcome contributions from the community! If you find an open LLM that works well with MemGPT, please open a PR with a model wrapper and we'll merge it ASAP.
+
+```python
+class Airoboros21Wrapper(LLMChatCompletionWrapper):
+    """Wrapper for Airoboros 70b v2.1: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1"""
+
+    def chat_completion_to_prompt(self, messages, functions):
+        """
+        Examples for how airoboros expects its prompt inputs: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1#prompt-format
+        Examples for how airoboros expects to see function schemas: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1#agentfunction-calling
+        """
+
+    def output_to_chat_completion_response(self, raw_llm_output):
+        """Turn raw LLM output into a ChatCompletion style response with:
+        "message" = {
+            "role": "assistant",
+            "content": ...,
+            "function_call": {
+                "name": ...
+                "arguments": {
+                    "arg1": val1,
+                    ...
+                }
+            }
+        }
+        """
+```
+See full file [here](llm_chat_completion_wrappers/airoboros.py). WebUI exposes a lot of parameters that can dramatically change LLM outputs, to change these you can modify the [WebUI settings file](/memgpt/local_llm/webui/settings.py).
+
+### Running the example
+
+```sh
+# running airoboros behind a textgen webui server
+export OPENAI_API_BASE = <pointing at webui server>
+export BACKEND_TYPE = webui
+
+# using --no_verify because this airoboros example does not output inner monologue, just functions
+# airoboros is able to properly call `send_message`
+$ python3 main.py --no_verify
+
+Running... [exit by typing '/exit']
+💭 Bootup sequence complete. Persona activated. Testing messaging functionality.
+
+💭 None
+🤖 Welcome! My name is Sam. How can I assist you today?
+Enter your message: My name is Brad, not Chad...
+
+💭 None
+⚡🧠 [function] updating memory with core_memory_replace:
+         First name: Chad
+        → First name: Brad
+```
+
+---
+
+## Status of ChatCompletion w/ function calling and open LLMs
+
+MemGPT uses function calling to do memory management. With [OpenAI's ChatCompletion API](https://platform.openai.com/docs/api-reference/chat/), you can pass in a function schema in the `functions` keyword arg, and the API response will include a `function_call` field that includes the function name and the function arguments (generated JSON). How this works under the hood is your `functions` keyword is combined with the `messages` and `system` to form one big string input to the transformer, and the output of the transformer is parsed to extract the JSON function call.
+
+In the future, more open LLMs and LLM servers (that can host OpenAI-compatable ChatCompletion endpoints) may start including parsing code to do this automatically as standard practice. However, in the meantime, when you see a model that says it supports “function calling”, like Airoboros, it doesn't mean that you can just load Airoboros into a ChatCompletion-compatable endpoint like WebUI, and then use the same OpenAI API call and it'll just work.
+
+1. When a model page says it supports function calling, they probably mean that the model was finetuned on some function call data (not that you can just use ChatCompletion with functions out-of-the-box). Remember, LLMs are just string-in-string-out, so there are many ways to format the function call data. E.g. Airoboros formats the function schema in YAML style (see https://huggingface.co/jondurbin/airoboros-l2-70b-3.1.2#agentfunction-calling) and the output is in JSON style. To get this to work behind a ChatCompletion API, you still have to do the parsing from `functions` keyword arg (containing the schema) to the model's expected schema style in the prompt (YAML for Airoboros), and you have to run some code to extract the function call (JSON for Airoboros) and package it cleanly as a `function_call` field in the response.
+
+2. Partly because of how complex it is to support function calling, most (all?) of the community projects that do OpenAI ChatCompletion endpoints for arbitrary open LLMs do not support function calling, because if they did, they would need to write model-specific parsing code for each one.
+
+## What is this all this extra code for?
+
+Because of the poor state of function calling support in existing ChatCompletion API serving code, we instead provide a light wrapper on top of ChatCompletion that adds parsers to handle function calling support. These parsers need to be specific to the model you're using (or at least specific to the way it was trained on function calling). We hope that our example code will help the community add additional compatability of MemGPT with more function-calling LLMs - we will also add more model support as we test more models and find those that work well enough to run MemGPT's function set.
+
+To run the example of MemGPT with Airoboros, you'll need to host the model behind some LLM web server (for example [webui](https://github.com/oobabooga/text-generation-webui#starting-the-web-ui)). Then, all you need to do is point MemGPT to this API endpoint by setting the environment variables `OPENAI_API_BASE` and `BACKEND_TYPE`. Now, instead of calling ChatCompletion on OpenAI's API, MemGPT will use it's own ChatCompletion wrapper that parses the system, messages, and function arguments into a format that Airoboros has been finetuned on, and once Airoboros generates a string output, MemGPT will parse the response to extract a potential function call (knowing what we know about Airoboros expected function call output).
--- a/memgpt/local_llm/init.py
+++ b/memgpt/local_llm/init.py
--- a/memgpt/local_llm/chat_completion_proxy.py
+++ b/memgpt/local_llm/chat_completion_proxy.py
@@ -0,0 +1,77 @@
+"""Key idea: create drop-in replacement for agent's ChatCompletion call that runs on an OpenLLM backend"""
+
+import os
+import requests
+import json
+
+from .webui.api import get_webui_completion
+from .llm_chat_completion_wrappers import airoboros
+from .utils import DotDict
+
+HOST = os.getenv("OPENAI_API_BASE")
+HOST_TYPE = os.getenv("BACKEND_TYPE")  # default None == ChatCompletion
+DEBUG = False
+
+
+async def get_chat_completion(
+    model,  # no model, since the model is fixed to whatever you set in your own backend
+    messages,
+    functions,
+    function_call="auto",
+):
+    if function_call != "auto":
+        raise ValueError(f"function_call == {function_call} not supported (auto only)")
+
+    if model == "airoboros_v2.1":
+        llm_wrapper = airoboros.Airoboros21Wrapper()
+    else:
+        # Warn the user that we're using the fallback
+        print(
+            f"Warning: could not find an LLM wrapper for {model}, using the airoboros wrapper"
+        )
+        llm_wrapper = airoboros.Airoboros21Wrapper()
+
+    # First step: turn the message sequence into a prompt that the model expects
+    prompt = llm_wrapper.chat_completion_to_prompt(messages, functions)
+    if DEBUG:
+        print(prompt)
+
+    try:
+        if HOST_TYPE == "webui":
+            result = get_webui_completion(prompt)
+        else:
+            print(f"Warning: BACKEND_TYPE was not set, defaulting to webui")
+            result = get_webui_completion(prompt)
+    except requests.exceptions.ConnectionError as e:
+        raise ValueError(f"Was unable to connect to host {HOST}")
+
+    if result is None or result == "":
+        raise Exception(f"Got back an empty response string from {HOST}")
+
+    chat_completion_result = llm_wrapper.output_to_chat_completion_response(result)
+    if DEBUG:
+        print(json.dumps(chat_completion_result, indent=2))
+
+    # unpack with response.choices[0].message.content
+    response = DotDict(
+        {
+            "model": None,
+            "choices": [
+                DotDict(
+                    {
+                        "message": DotDict(chat_completion_result),
+                        "finish_reason": "stop",  # TODO vary based on backend response
+                    }
+                )
+            ],
+            "usage": DotDict(
+                {
+                    # TODO fix, actually use real info
+                    "prompt_tokens": 0,
+                    "completion_tokens": 0,
+                    "total_tokens": 0,
+                }
+            ),
+        }
+    )
+    return response
--- a/memgpt/local_llm/llm_chat_completion_wrappers/init.py
+++ b/memgpt/local_llm/llm_chat_completion_wrappers/init.py
--- a/memgpt/local_llm/llm_chat_completion_wrappers/airoboros.py
+++ b/memgpt/local_llm/llm_chat_completion_wrappers/airoboros.py
@@ -0,0 +1,204 @@
+import json
+
+from .wrapper_base import LLMChatCompletionWrapper
+
+
+class Airoboros21Wrapper(LLMChatCompletionWrapper):
+    """Wrapper for Airoboros 70b v2.1: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1
+
+    Note: this wrapper formats a prompt that only generates JSON, no inner thoughts
+    """
+
+    def __init__(
+        self,
+        simplify_json_content=True,
+        clean_function_args=True,
+        include_assistant_prefix=True,
+        include_opening_brace_in_prefix=True,
+        include_section_separators=True,
+    ):
+        self.simplify_json_content = simplify_json_content
+        self.clean_func_args = clean_function_args
+        self.include_assistant_prefix = include_assistant_prefix
+        self.include_opening_brance_in_prefix = include_opening_brace_in_prefix
+        self.include_section_separators = include_section_separators
+
+    def chat_completion_to_prompt(self, messages, functions):
+        """Example for airoboros: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1#prompt-format
+
+        A chat.
+        USER: {prompt}
+        ASSISTANT:
+
+        Functions support: https://huggingface.co/jondurbin/airoboros-l2-70b-2.1#agentfunction-calling
+
+            As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
+
+            Input: I want to know how many times 'Python' is mentioned in my text file.
+
+            Available functions:
+            file_analytics:
+              description: This tool performs various operations on a text file.
+              params:
+                action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
+                filters:
+                  keyword: The word or phrase we want to search for.
+
+        OpenAI functions schema style:
+
+            {
+                "name": "send_message",
+                "description": "Sends a message to the human user",
+                "parameters": {
+                    "type": "object",
+                    "properties": {
+                        # https://json-schema.org/understanding-json-schema/reference/array.html
+                        "message": {
+                            "type": "string",
+                            "description": "Message contents. All unicode (including emojis) are supported.",
+                        },
+                    },
+                    "required": ["message"],
+                }
+            },
+        """
+        prompt = ""
+
+        # System insturctions go first
+        assert messages[0]["role"] == "system"
+        prompt += messages[0]["content"]
+
+        # Next is the functions preamble
+        def create_function_description(schema):
+            # airorobos style
+            func_str = ""
+            func_str += f"{schema['name']}:"
+            func_str += f"\n  description: {schema['description']}"
+            func_str += f"\n  params:"
+            for param_k, param_v in schema["parameters"]["properties"].items():
+                # TODO we're ignoring type
+                func_str += f"\n    {param_k}: {param_v['description']}"
+            # TODO we're ignoring schema['parameters']['required']
+            return func_str
+
+        # prompt += f"\nPlease select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format."
+        prompt += f"\nPlease select the most suitable function and parameters from the list of available functions below, based on the ongoing conversation. Provide your response in JSON format."
+        prompt += f"\nAvailable functions:"
+        for function_dict in functions:
+            prompt += f"\n{create_function_description(function_dict)}"
+
+        def create_function_call(function_call):
+            """Go from ChatCompletion to Airoboros style function trace (in prompt)
+
+            ChatCompletion data (inside message['function_call']):
+                "function_call": {
+                    "name": ...
+                    "arguments": {
+                        "arg1": val1,
+                        ...
+                    }
+
+            Airoboros output:
+                {
+                  "function": "send_message",
+                  "params": {
+                    "message": "Hello there! I am Sam, an AI developed by Liminal Corp. How can I assist you today?"
+                  }
+                }
+            """
+            airo_func_call = {
+                "function": function_call["name"],
+                "params": json.loads(function_call["arguments"]),
+            }
+            return json.dumps(airo_func_call, indent=2)
+
+        # Add a sep for the conversation
+        if self.include_section_separators:
+            prompt += "\n### INPUT"
+
+        # Last are the user/assistant messages
+        for message in messages[1:]:
+            assert message["role"] in ["user", "assistant", "function"], message
+
+            if message["role"] == "user":
+                if self.simplify_json_content:
+                    try:
+                        content_json = json.loads(message["content"])
+                        content_simple = content_json["message"]
+                        prompt += f"\nUSER: {content_simple}"
+                    except:
+                        prompt += f"\nUSER: {message['content']}"
+            elif message["role"] == "assistant":
+                prompt += f"\nASSISTANT: {message['content']}"
+                # need to add the function call if there was one
+                if message["function_call"]:
+                    prompt += f"\n{create_function_call(message['function_call'])}"
+            elif message["role"] == "function":
+                # TODO find a good way to add this
+                # prompt += f"\nASSISTANT: (function return) {message['content']}"
+                prompt += f"\nFUNCTION RETURN: {message['content']}"
+                continue
+            else:
+                raise ValueError(message)
+
+        # Add a sep for the response
+        if self.include_section_separators:
+            prompt += "\n### RESPONSE"
+
+        if self.include_assistant_prefix:
+            prompt += f"\nASSISTANT:"
+            if self.include_opening_brance_in_prefix:
+                prompt += "\n{"
+
+        return prompt
+
+    def clean_function_args(self, function_name, function_args):
+        """Some basic MemGPT-specific cleaning of function args"""
+        cleaned_function_name = function_name
+        cleaned_function_args = function_args.copy()
+
+        if function_name == "send_message":
+            # strip request_heartbeat
+            cleaned_function_args.pop("request_heartbeat", None)
+
+        # TODO more cleaning to fix errors LLM makes
+        return cleaned_function_name, cleaned_function_args
+
+    def output_to_chat_completion_response(self, raw_llm_output):
+        """Turn raw LLM output into a ChatCompletion style response with:
+        "message" = {
+            "role": "assistant",
+            "content": ...,
+            "function_call": {
+                "name": ...
+                "arguments": {
+                    "arg1": val1,
+                    ...
+                }
+            }
+        }
+        """
+        if self.include_opening_brance_in_prefix and raw_llm_output[0] != "{":
+            raw_llm_output = "{" + raw_llm_output
+
+        try:
+            function_json_output = json.loads(raw_llm_output)
+        except Exception as e:
+            raise Exception(f"Failed to decode JSON from LLM output:\n{raw_llm_output}")
+        function_name = function_json_output["function"]
+        function_parameters = function_json_output["params"]
+
+        if self.clean_func_args:
+            function_name, function_parameters = self.clean_function_args(
+                function_name, function_parameters
+            )
+
+        message = {
+            "role": "assistant",
+            "content": None,
+            "function_call": {
+                "name": function_name,
+                "arguments": json.dumps(function_parameters),
+            },
+        }
+        return message
--- a/memgpt/local_llm/llm_chat_completion_wrappers/wrapper_base.py
+++ b/memgpt/local_llm/llm_chat_completion_wrappers/wrapper_base.py
@@ -0,0 +1,13 @@
+from abc import ABC, abstractmethod
+
+
+class LLMChatCompletionWrapper(ABC):
+    @abstractmethod
+    def chat_completion_to_prompt(self, messages, functions):
+        """Go from ChatCompletion to a single prompt string"""
+        pass
+
+    @abstractmethod
+    def output_to_chat_completion_response(self, raw_llm_output):
+        """Turn the LLM output string into a ChatCompletion response"""
+        pass
--- a/memgpt/local_llm/utils.py
+++ b/memgpt/local_llm/utils.py
@@ -0,0 +1,8 @@
+class DotDict(dict):
+    """Allow dot access on properties similar to OpenAI response object"""
+
+    def __getattr__(self, attr):
+        return self.get(attr)
+
+    def __setattr__(self, key, value):
+        self[key] = value
--- a/memgpt/local_llm/webui/api.py
+++ b/memgpt/local_llm/webui/api.py
@@ -0,0 +1,33 @@
+import os
+import requests
+
+from .settings import SIMPLE
+
+HOST = os.getenv("OPENAI_API_BASE")
+HOST_TYPE = os.getenv("BACKEND_TYPE")  # default None == ChatCompletion
+WEBUI_API_SUFFIX = "/api/v1/generate"
+DEBUG = False
+
+
+def get_webui_completion(prompt, settings=SIMPLE):
+    """See https://github.com/oobabooga/text-generation-webui for instructions on how to run the LLM web server"""
+
+    # Settings for the generation, includes the prompt + stop tokens, max length, etc
+    request = settings
+    request["prompt"] = prompt
+
+    try:
+        URI = f"{HOST.strip('/')}{WEBUI_API_SUFFIX}"
+        response = requests.post(URI, json=request)
+        if response.status_code == 200:
+            result = response.json()
+            result = result["results"][0]["text"]
+            if DEBUG:
+                print(f"json API response.text: {result}")
+        else:
+            raise Exception(f"API call got non-200 response code for address: {URI}")
+    except:
+        # TODO handle gracefully
+        raise
+
+    return result
--- a/memgpt/local_llm/webui/settings.py
+++ b/memgpt/local_llm/webui/settings.py
@@ -0,0 +1,12 @@
+SIMPLE = {
+    "stopping_strings": [
+        "\nUSER:",
+        "\nASSISTANT:",
+        # '\n' +
+        # '</s>',
+        # '<|',
+        # '\n#',
+        # '\n\n\n',
+    ],
+    "truncation_length": 4096,  # assuming llama2 models
+}
--- a/memgpt/openai_tools.py
+++ b/memgpt/openai_tools.py
@@ -3,8 +3,16 @@ import random
 import os
 import time

+from .local_llm.chat_completion_proxy import get_chat_completion
+
+HOST = os.getenv("OPENAI_API_BASE")
+HOST_TYPE = os.getenv("BACKEND_TYPE")  # default None == ChatCompletion
+
 import openai

+if HOST is not None:
+    openai.api_base = HOST
+

 def retry_with_exponential_backoff(
    func,
@@ -102,18 +110,24 @@ def aretry_with_exponential_backoff(

@aretry_with_exponential_backoff
 async def acompletions_with_backoff(**kwargs):
-    azure_openai_deployment = os.getenv('AZURE_OPENAI_DEPLOYMENT')
-    if azure_openai_deployment is not None:
-        kwargs['deployment_id'] = azure_openai_deployment
-    return await openai.ChatCompletion.acreate(**kwargs)
+    # Local model
+    if HOST_TYPE is not None:
+        return await get_chat_completion(**kwargs)
+
+    # OpenAI / Azure model
+    else:
+        azure_openai_deployment = os.getenv("AZURE_OPENAI_DEPLOYMENT")
+        if azure_openai_deployment is not None:
+            kwargs["deployment_id"] = azure_openai_deployment
+        return await openai.ChatCompletion.acreate(**kwargs)


@aretry_with_exponential_backoff
 async def acreate_embedding_with_backoff(**kwargs):
    """Wrapper around Embedding.acreate w/ backoff"""
-    azure_openai_deployment = os.getenv('AZURE_OPENAI_DEPLOYMENT')
+    azure_openai_deployment = os.getenv("AZURE_OPENAI_DEPLOYMENT")
    if azure_openai_deployment is not None:
-        kwargs['deployment_id'] = azure_openai_deployment
+        kwargs["deployment_id"] = azure_openai_deployment
    return await openai.Embedding.acreate(**kwargs)


@@ -121,6 +135,6 @@ async def async_get_embedding_with_backoff(text, model="text-embedding-ada-002")
    """To get text embeddings, import/call this function
    It specifies defaults + handles rate-limiting + is async"""
    text = text.replace("\n", " ")
-    response = await acreate_embedding_with_backoff(input = [text], model=model)
-    embedding = response['data'][0]['embedding']
-    return embedding
+    response = await acreate_embedding_with_backoff(input=[text], model=model)
+    embedding = response["data"][0]["embedding"]
+    return embedding
--- a/requirements.txt
+++ b/requirements.txt
@@ -9,6 +9,7 @@ pybars3
 pymupdf
 python-dotenv
 pytz
+questionary
 rich
 tiktoken
 timezonefinder