TypeAgent allows you to create various types of LLM agents,agent types include Memgpt, OpenDevin, SWE, AIOS,Concordia,Generative Agents etc, every type of agent has a unique architecture which has unique functionality
TypeAgent makes it easy to build and deploy stateful,various types of LLM agents with support for:
- Various types of LLM agents,Integrated in unified TypeAgent framework, seamless switching various types of llm agents.
- LLM Type-architecture-functionality architecture which ensure every type of agent has unique feature
- Long term memory/state managemet
- Basic RAG workflow for knowledge base which created by external data sources (e.g. PDF files)
- Intelligent switching knowledge base
- Defining and calling custom tools
You can also use TypeAgent to depoy agents as a service. You can use a TypeAgent server to run a multi-user, multi-agent application on top of supported type of agents and supported LLM providers.
Install TypeAgent:
pip install -U pytypeagent
You can create and chat with a TypeAgent agent by running typeagent run
in your CLI. The run
command supports the following optional flags (see the [CLI documentation] for the full list of flags):
--agent
: (str) Name of agent to create or to resume chatting with.--type
:(str) type of agent to create or to resume chatting with,default type is memgpt--first
: (str) Allow user to sent the first message.--debug
: (bool) Show debug logs (default=False)
You can view the list of available in-chat commands (e.g. /memory
, /exit
) in the [CLI documentation].
1: Run with the CLI:
- Run
typeagent configure
- Run
typeagent server
- Go to
localhost:8283
in the browser to view the developer portal
Once the server is running, you can use the [Python client] or [REST API] to connect to localhost:8283
(if you're running with the CLI) to create users, agents, and more. The service requires authentication with a TypeAgent admin password, which can be set with running export TYPEAGENT_SERVER_PASS=password
.
TypeAgent is designed to be Agent Type agnostic. The following Agent Type are supported:
Type | support status |
---|---|
Memgpt | ✅ |
OpenDevin_Planer | ✅ |
OpenDevin_SWE | ❌ |
OpenDevin_CodeAct | ❌ |
OpenDevin_Micro | ❌ |
OpenDevin_Monologue | ❌ |
Devika | ❌ |
Perplexica | ❌ |
Concordia | ❌ |
Charlie Mnemonic | ❌ |
AIOS | ❌ |
Open Interpreter | ❌ |
SalesGPT | ❌ |
SWE | ❌ |
Generative Agents | ❌ |
The implementation methods of all the above types of agents are different. Now they are integrated into the unified interface through TypeAgent.
Later, different types of agents will be added one after another
TypeAgent is designed to be model and provider agnostic. The following LLM and embedding endpoints are supported:
Provider | LLM Endpoint | Embedding Endpoint |
---|---|---|
OpenAI | ✅ | ✅ |
Azure OpenAI | ✅ | ✅ |
Google AI (Gemini) | ✅ | ❌ |
Anthropic (Claude) | ✅ | ❌ |
Groq | ✅ (alpha release) | ❌ |
Cohere API | ✅ | ❌ |
vLLM | ✅ | ❌ |
Ollama | ✅ | ❌ |
LM Studio | ✅ | ❌ |
koboldcpp | ✅ | ❌ |
oobabooga web UI | ✅ | ❌ |
llama.cpp | ✅ | ❌ |
HuggingFace TEI | ❌ | ✅ |
When using TypeAgent with open LLMs (such as those downloaded from HuggingFace), the performance of TypeAgent will be highly dependent on the LLM's function calling ability ,Language understanding and reasoning skills.
See full documentation at:Coming soon
-
If you are interested in the project ,you can make a issue, i am working on that .......
-
Our codebase for the TypeAgent builds heavily on MemGPT codebase Thanks for open-sourcing! ,Our code will be open soon.....
-
The implementation of the all type of agent refer to respective source code,Thanks for open-sourcing! and integrated into MemGPT structure
By using TypeAgent and related TypeAgent services (such as the TypeAgent endpoint ), you agree to our privacy policy and terms of service.
You can view (and comment on!) the TypeAgent developer roadmap on GitHub: Coming soon