Repolode
Submit
framerslab

framerslab/agentos

TypeScript AI agent framework with persistent cognitive memory, optional HEXACO personality, multi-agent orchestration, and a unified interface across 13 LLM providers.

View on GitHubWebsite
Stars
681
Forks
100
Open issues
24
Last commit
12 hours ago

What is it?

AgentOS is an open-source TypeScript framework for building AI agents that keep memory across sessions, adapt their behavior, and can create new tools at runtime. A basic agent is created with a single call that picks a provider from configuration, and sessions keep conversation history that later requests replay.

The framework includes a cognitive memory layer with Ebbinghaus-style decay and several memory mechanisms such as reconsolidation and retrieval-induced forgetting. Agents can carry an optional HEXACO personality vector that shapes prompts and memory encoding. Multi-agent teams can run under six coordination strategies, including sequential, parallel, debate, review-loop, hierarchical and graph, and the package also provides guardrail packs, voice pipeline support and RAG retrieval over seven vector backends.

AgentOS is split into a core package plus separate extension, skill and registry packages. The project publishes benchmark results on the LongMemEval dataset with its methodology in a companion repository.

Key features

  • 01Persistent cognitive memory with decay and several memory mechanisms.
  • 02Optional HEXACO personality traits that shape prompts and memory.
  • 03Multi-agent teams with six coordination strategies.
  • 04One interface across 13 LLM providers, with API key pooling and fallback.
  • 05Runtime tool generation, with generated tools reviewed by an LLM judge before use.
  • 06Guardrail packs, voice support and RAG over seven vector backends.

Use cases

  • Building chatbots or tutors that remember users across multiple conversations.
  • Orchestrating several specialized agents, such as researcher, writer and reviewer, on one task.
  • Prototyping agents that work across different LLM providers without rewriting code.
  • Adding voice or chat-channel interfaces to an agent backend.

Quick start

  1. 1

    Install the package

    npm install @framers/agentos
  2. 2

    Create an agent and send a message (see README example; set a provider API key in your environment)

Things to know

  • Cognitive memory benchmark numbers come from the project's own agentos-bench repository, not independent verification.
  • Some features such as the full GMI runtime, self-modification tools and guardrails require the AgentOS runtime or specific options.
  • Runtime tool forging runs code in node:vm by default, which the README states is not a security boundary.

Topics

#agent-framework#typescript#llm#agent-memory#multi-agent#rag#tool-use#voice-ai#agentic-ai#ai-agent-framework

FAQ

What is agentos?
TypeScript AI agent framework with persistent cognitive memory, optional HEXACO personality, multi-agent orchestration, and a unified interface across 13 LLM providers. AgentOS is an open-source TypeScript framework for building AI agents that keep memory across sessions, adapt their behavior, and can create new tools at runtime. A basic agent is created with a single call that picks a provider from configuration, and sessions keep conversation history that later requests replay.
Is agentos free to use? What license does it have?
agentos is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is agentos still maintained?
Actively maintained. The last commit to agentos was on Oct 19, 2026.
What can I use agentos for?
Building chatbots or tutors that remember users across multiple conversations. Orchestrating several specialized agents, such as researcher, writer and reviewer, on one task. Prototyping agents that work across different LLM providers without rewriting code. Adding voice or chat-channel interfaces to an agent backend.
How do I install agentos?
Install the package: npm install @framers/agentos Create an agent and send a message (see README example; set a provider API key in your environment)
Where can I get agentos?
The source code and installation instructions are on GitHub: https://github.com/framerslab/agentos

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects