Knowledge Operating System for AI Agents
Atlas separates knowledge from reasoning. Compile verifiable,
deterministic decision trees into a portable .atlas binary — so your agents make the
right call without an LLM call. Built for people and the agents working alongside them.
LangChain, CrewAI, AutoGen, smolagents — every LLM-first framework shares the same flaw: knowledge lives in the context window.
No structure, no persistence, no verification. The same question gets a different answer every run.
Each agent step is an LLM call. At $0.10–1.00+ per decision, production agents bleed cash on questions with known answers.
When the model hallucinates there's no safety net — no verification, no audit trail, no replay.
EU AI Act obligations demand you explain why an agent did what it did. Prompt-only systems can't.
Atlas is an operating system for engineering knowledge — not for computers, but for the agents that reason on top of them.
$ atlas init my-package --template flutter
Created knowledge package 'my-package'
$ atlas compile my-package.md decisions/ --out my-package.atlas
Compiled 12 nodes, 8 edges, 1 decision tree → my-package.atlas
$ atlas solve --bundle my-package.atlas "which widget?"
Matched 5 nodes. Confidence: 0.95
$ atlas decide --bundle my-package.atlas "widget" -c "needs_state=true"
Path: widget_state → stateful_vs_builder → stateful_builder
Rationale: Use StatefulWidget with a builder pattern...
$ |
.atlas binary, memory-mapped, signedConcepts, APIs, workflows — all in a simple, human-readable format.
.atlas binaryZstd-compressed, memory-mapped, blake3-hashed. Portable across any platform.
Python SDK, MCP server, CLI, or Studio. Any interface. Zero LLM calls for 80% of queries.
Share, discover, and reuse knowledge packages. Like a container registry, but for agent knowledge. 51 packages and growing.
SSO, GDPR, audit, RBAC — all included. No paywalls, no enterprise license fees.
Authenticate with Google, Azure AD, Okta, or any OpenID Connect provider. JWT verification via JWKS.
DSAR, data portability export, and right to erasure endpoints. Built for SOC 2 and EU AI Act readiness.
Daily R2 backups with 30-day retention. Manual trigger endpoint for on-demand snapshots.
Role-based access (admin, publisher, viewer). Full audit trail with timestamps, IPs, and API key tracking.
60 requests/minute per API key with X-RateLimit-Reset headers. Protection on all endpoints.
Apache 2.0 license. Audit the code, fork it, deploy it yourself. No enterprise paywall — ever.
Also available: cargo install atlas-cli · VS Code Extension · GitHub Action · MCP Server
| Atlas | smolagents | LangChain | CrewAI | AutoGen | |
|---|---|---|---|---|---|
| Deterministic decisions | Yes LLM-free | No | No | No | No |
| Built-in verification | Yes | No | Manual | No | No |
| Knowledge composition | Yes | No | RAG | No | No |
| Offline / Air-gapped | Yes | No | Partial | No | No |
| Audit trail | Yes Crypto | No | LangSmith | No | No |
| SSO / OIDC | Yes | No | No | No | No |
| GDPR compliance | Yes | No | No | No | No |
| Cost per query | ~$0.0001 | $0.10–1.00 | $0.10–1.00 | $0.10–1.00 | $0.10–1.00 |
| EU AI Act ready | Yes Built-in | No | No | No | No |
| Open source | Yes | Yes | Yes | Yes | Yes |