LLM

LLM documentation pack.

Raw, complete, copy-paste documentation for AI assistants. One click gives an LLM everything it needs to build with Actae — no HTML, no navigation, just the facts.

Why a pack?

LLMs consume plain text best. The pack is the exact same documentation the Actae team uses for AI-assisted development — curated, public-safe markdown with every internal anchor stripped. Paste it into any chat and the model knows the product surface: features, routes, JSON fields, defaults, failure semantics, SDKs, adapters, and testing.

Copy everything

The full pack concatenates all 15 references into one file (242 KB, ~5,205 lines). For most conversations this is the one to grab.

Prefer a file?

Every artifact below is also served as a plain file — fetch it with curl, download it, or reference it by URL: /llm/actae-pack.md

Per-topic dumps

When a conversation only needs one area, copy the matching reference instead of the whole pack.

ReferenceContentsCopy
Framework adapters382 lines · 16 KB
AgentSession lifecycle313 lines · 10 KB
Coordination: consumer groups and wake-ups164 lines · 7 KB
Counterfactual debugging103 lines · 5 KB
Durable execution48 lines · 3 KB
Execution integrity58 lines · 2 KB
Experiments85 lines · 4 KB
Actae feature reference965 lines · 50 KB
Fork & resume workflow559 lines · 26 KB
Fork runtime429 lines · 20 KB
Actae Go SDK reference238 lines · 12 KB
Actae Python SDK reference1,024 lines · 43 KB
State system236 lines · 10 KB
Idempotent tool executions238 lines · 12 KB
Actae TypeScript SDK reference363 lines · 20 KB

How it stays current

The pack is regenerated from the repository documentation on every build (npm run sync-llm-docs). The curated manifest includes only public-safe references; internal runbooks, plans, and server internals are excluded by design.

Start with the quickstart, then fork, coordinate, and measure — with the pack in your chat, the LLM is never guessing about Actae.