a recursive self-improving harness designed to help your agents (and future iterations of those agents) succeed on any task

autocontext is a harness for agent improvement. Give it a goal, it runs the task against evaluation, keeps the useful lessons, discards dead ends, and leaves traces, reports, playbooks, datasets, and optional local-model training artifacts for the next run.

Docs: autocontext.ai/docs · quickstart · CLI reference · changelog

Install

The PyPI package is autocontext; the CLI is autoctx. The npm package is autoctx (not the unrelated autocontext npm package). Provider variables live in .env.example.

30-Second Run

Pi is the lowest-friction provider because it uses your local agent auth:

AUTOCONTEXT_AGENT_PROVIDER=pi \
AUTOCONTEXT_PI_COMMAND=pi \
autoctx solve "improve customer-support replies for billing disputes" --iterations 3

Use AUTOCONTEXT_AGENT_PROVIDER=anthropic, openai-compatible, openrouter, claude-cli, codex, pi-rpc, or another provider when you need that runtime. See agent integration for the full matrix.

Running it on your own GPU instead? Self-hosted models covers the whole loop on vLLM, Ollama, or any OpenAI-compatible endpoint — including what each role actually resolves to, and why constrained output matters more on open weights. Self-hosted endpoints can additionally declare AUTOCONTEXT_PROVIDER_HOSTING=local and a fast, mid_tier, or frontier AUTOCONTEXT_PROVIDER_CAPABILITY; role-specific endpoints use matching <ROLE>_PROVIDER_* declarations.

Agent Entry Points

• Pi: install pi-autocontext, then ask Pi to solve, judge, improve, list, or inspect runs through the packaged skill.
• MCP clients: run autoctx mcp-serve or bunx autoctx mcp-serve and expose the tools to Claude Code, Cursor, or another MCP client.
• Hermes: export the CLI-first skill with uv run autoctx hermes export-skill --with-references --json.

Full setup: autocontext/docs/agent-integration.md.

What A Run Leaves Behind

runs/<run_id>/
├── trace.jsonl
├── generations/<n>/{strategy.json,analysis.md,score.json}
├── report.md
└── artifacts/

knowledge/<scenario>/
├── playbook.md
├── hints.md
└── tools/

Everything is filesystem-first: inspect it, diff it, replay it, export it, or feed it into training.

Core Surfaces

Python owns the full control-plane package; TypeScript owns several operator-facing surfaces, the TUI, and Node runtime adapters. Start with autocontext/README.md or ts/README.md.

What's New in 0.14.0

• Live TypeScript task plans: interactive runs now advertise agent_task_plan_v1 and emit semantic task_plan_updated snapshots for initial planning, meaningful progress, replanning, completion, failure, and safe stop.
• Durable, privacy-safe replay: task-plan snapshots use stable identity and monotonic revisions, redact credential-shaped values, reject malformed or oversized frames atomically, and restore exactly across reconnects and server restarts. Python producer parity remains explicitly deferred.

Scenario Families

The shipped families cover games, agent tasks, simulations, artifact editing, investigations, workflows, negotiation, schema evolution, tool fragility, operator loops, and coordination. Python and TypeScript share the family vocabulary; see docs/scenario-parity-matrix.md for parity details.

Package Guides

Project Signals

Acknowledgments

Thanks to George for generously donating the autocontext name on PyPI.