OSS TanbouOn this page
Overview
AgentOS is a TypeScript framework that combines session history, cognitive memory, tool calling, runtime tool forging, and multi-agent coordination in one runtime. Multiple LLM providers, guardrails, and voice capabilities share the same framework.
Features and best fit
Based on official documentation; not hands-on tested · Content checked:
Key features
Combine session history with cognitive memory
Agents can use conversation history plus episodic, semantic, and related memory mechanisms to recover information across longer sessions.
Generate, approve, and reuse tools at runtime
When the static toolkit is insufficient, an agent can generate a TypeScript function and Zod schema, pass an LLM-judge approval step, and add the sandboxed tool to the session catalog.
Sources: [2]
Best fit
Fits teams consolidating memory, tools, and multi-agent logic in TypeScript
It is useful when provider routing, agent teams, guardrails, voice, and memory should share one runtime instead of being assembled from unrelated libraries.
Sources: [2]
Before adoption
Manage provider credentials and model versions explicitly
Providers can be auto-detected from environment variables, but production systems should define provider priority or defaults and pin model IDs to avoid unexpected selection changes.
Sources: [3]
Runtime-generated tools still need least-privilege boundaries
Tool forging uses a constrained node:vm sandbox, but any external credentials or network access exposed to generated tools should remain deliberately minimal.
Sources: [2]
Official sources
- [1]framerslab/agentos repository(2026-09-30)
- [2]AgentOS README(2026-09-30)
- [3]AgentOS getting started guide(2026-09-30)
- [4]AgentOS Apache-2.0 license(2026-09-30)
- [5]AgentOS v0.10.22 release(2026-09-30)
Supplemental curator note
The requested framersai/agentos repository now resolves to framerslab/agentos. AgentOS is an Apache-2.0 TypeScript framework combining session and cognitive memory, runtime tool forging, multi-agent orchestration, guardrails, and voice capabilities. Provider credentials can come from environment variables, inline configuration, or a module-level default; production systems should avoid embedding API keys in source code.
Try it in 3 steps
- 1
Install AgentOS 0.10.22
Create a minimal Node.js project with the package pinned to the release version.
mkdir agentos-quickstart && cd agentos-quickstart && npm init -y && npm install @framers/agentos@0.10.22 - 2
Set a provider credential
Replace the placeholder with your provider credential and keep API keys out of source control.
export OPENAI_API_KEY=YOUR_KEY - 3
Run a minimal generateText call
Verify package operation and provider connectivity with a single LLM call.
node --input-type=module -e "import { generateText } from '@framers/agentos'; const { text } = await generateText({ provider: 'openai', apiKey: process.env.OPENAI_API_KEY, prompt: 'Reply in one sentence: AgentOS is running.' }); console.log(text)"
Growth
Growth trends · Last 30 days
672 Stars
Trend data is still being collected.
Built with
Categories and tags
GitHub data
GitHub dataView detailed GitHub data
GitHub Topics
- agent-framework
- agent-memory
- agentic-ai
- ai-agent-framework
- ai-agents
- autonomous-agents
- cognitive-memory
- emergent-behavior
- guardrails
- hexaco
- llm
- llm-orchestration
- Stars
- 672
- Forks
- 98
- Watchers
- 2
- Open issues
- 11
- Owner type
- Organization
- Primary language
- TypeScript
- License
- Apache-2.0
- Repository last updated
- Sep 30, 2026
Related information
Write a related articleShare a guide or use case for this OSS in Markdown. Articles are published after administrator approval.
Explore next
- Heliox-OS65 Stars
3 shared tag(s)
Control an existing OS through natural language, voice, and gestures while routing real actions through permissions, approvals, and verification
Python - OpenClaw390,815 Stars
2 shared tag(s) · 2 shared category(s) · Same language
Connect a persistent assistant to channels, tools, memory, and device nodes through a local Gateway control plane
TypeScript - NanoClaw30,865 Stars
2 shared tag(s) · 2 shared category(s) · Same language
Isolate AI agents in per-group containers while connecting messaging, memory, schedules and credential gateways
TypeScript - FastGPT29,773 Stars
2 shared tag(s) · 2 shared category(s) · Same language
assemble document-grounded answers and AI workflows visually
TypeScript - Budibase28,327 Stars
2 shared tag(s) · 2 shared category(s) · Same language
Build internal apps, automations, and AI agents visually on top of data sources and approval policies
TypeScript - Activepieces24,794 Stars
2 shared tag(s) · 2 shared category(s) · Same language
put AI, MCP, and human decisions in the same workflow
TypeScript
Report incorrect information
Tell us if any listing information is incorrect or outdated.