Causal

Build agents that learn from experience

Make your agents smarter and more efficient with experience memory. Record what agents learned, what they did, which outcomes followed and what they should change next time.

Join waitlist Schedule demo
Get started
Works with
ClaudeCursorCodexCopilotGeminiOpenCodeOpenclawHermesTypeScriptPython

Share experience between agents

Causal shares experience and context between your agents, so you spend no more time or tokens re-explaining yourself.

Every decision
auditable

Ask any agent why it acted, and Causal rebuilds the trail — the precedent it was served, what superseded what, and the signed source events underneath, provable to an auditor.

Anti-poisoning built in

A poisoned write competes with what your agents already learned. The experience with more evidence behind it wins, and the poison is tainted.

Start in one command

Works directly with Claude Code, Cursor, Codex, GitHub Copilot, Gemini CLI, OpenCode, OpenClaw, and Hermes. Anything outside that list uses the SDK.

$npx causal --runtimes=claude
  • 2 skills → ~/.claude/skills
  • MCP → ~/.claude.json (wired)
  • rule → ~/.claude/CLAUDE.md (appended)

Connect and learn.

Ingest experience from the tools and workflows you already run, and store it as memories your agents can use.

Start compounding

Your company brain

Causal keeps every team’s experience in one graph, scoped to each role. It compounds as the company works, and it stays when the people who built it leave.

See the enterprise story

Memory with enterprise-grade security

A retro arcade control grid routing a beam through a prism into ordered lanes

100% in your control

We never see your data. Delete any memory, or your whole account, whenever you want.

A retro-futurist vault with a bright circular center and spectral arches
GDPRCCPAHIPAA

Encrypted and provable

Your memory is encrypted at rest and in transit. Every write is signed, so tampering is provable. Enterprise runs under your own keys, in your own VPC.

A spectral proof chain leaving a crystalline archive and continuing toward the horizon

Your data stays yours

We never sell your data or train models on it. Export the whole graph in an open format and verify it offline.

Frequently Asked

What is Causal?

Causal is experience memory for AI agents. It records what your agents tried, what happened, and what to change next time, then serves that experience into the next decision. Each agent gets smarter and more efficient the more you use it.

How is Causal different from other agent memory?

Most memory tools store facts: what was said, embedded and fetched back by similarity. Causal stores experiences: cause and effect over time. An agent gets back what a past run tried and the outcome it caused, so it acts on precedent instead of re-reading context.

What does it work with?

Claude Code, Cursor, Codex, GitHub Copilot, Gemini CLI, OpenCode, OpenClaw, and Hermes. All wired in one command with npx causal. Python and TypeScript SDKs cover anything outside that list.

How does it get smarter the more I use it?

Every run leaves its outcome behind, so the next decision starts from what already worked and what already failed. Repeated mistakes stop repeating, and good moves compound. The graph is pooled across your agents. What one learns, the next already knows.