bot0.ai

Resources

Community

These projects grew out of shipping bot0.ai— the multi-turn control plane, ledger discipline, and regression harness we needed in production. They are MIT-licensed, independent of the commercial product surface, and meant to be useful on their own stacks.

Feel free to contribute, fork, and take them forward. Issues and PRs on GitHub are welcome.

The core bot0.ai taxonomy, ontology, and engines may also be made open source at some point, if and when that makes sense for the community and the product. This page will list them when they land.

Conversation Control Plane

conversation-control-plane

View on GitHub

A durable turn-ownership ledger for multi-agent chat — portable across how each turn is run (LangGraph, agent SDKs, Temporal, plain code, or a human). Not a replacement for those leaves; compose with them.

From the project README

Most orchestration stacks already track state (graph checkpoints, crew context, Temporal workflows, hosted chat sessions). What usually stays coupled to one runtime is the meaning of that state: who is foreground, what is pinned, and how resume/handoff work across specialists.

This package makes conversational authority an independent contract: single-writer decide_turn, thin projection (L1) + event journal (L2), COMPLETE ≠ ABANDON, multi-task suspend/resume. Semantic interpretation stays with the LLM; who owns the thread is code-owned. Persistence is assumed — DB-backed is not the USP; portable authority semantics are.

  • claim → decide_turn → handle → apply → release host cycle
  • Cross-runtime authority (same law if the run leaf changes)
  • HITL-ready · multi-task foreground · KindSpec + thin pins

https://github.com/walidnegm/conversation-control-plane

Conjecture Behaviour Runner

conjecture-behaviour-runner

View on GitHub

Catch state-law breaks that still look fine in chat — wrong exclusive owner, dropped pin, illegal restart — under pinned cognition for CI.

From the project README

In multi-turn agents, the reply can sound correct while the conversation machine underneath is wrong. Those are ledger and handoff bugs, not bad writing. Unit checks and LLM-as-judge evals usually miss them.

Principle: LLM proposes · code enforces. Conjecture regression-tests the enforce half after each scripted turn (owner · pin · handoff). Owners, kinds, and pins are host vocabulary — not a fixed Conjecture catalog. Pair with separate classifier tests for cognition drift; use Conjecture for “given these labels, did code still seal the ledger?”

  • Script · Turn · Driver · Observation · Invariant
  • Healthy-PASS / planted-FAIL discipline + CAQ-FM taxonomy
  • Ledger-store agnostic if Act projects owner/pins (DB, LangGraph, Temporal, …)

MIT · 0.1.6. Demos use sample kinds (e.g. cost_out) only as stand-ins — substitute your exclusive_owner and pin keys.

https://github.com/walidnegm/conjecture-behaviour-runner

Coming later (maybe)

Core bot0.ai taxonomy, ontology, and simulation engines are not open source today. If we publish them, they will appear here with the same honest framing: extracted from production use, useful without requiring the commercial platform, and open to contribution when the contracts are ready.