Ollama-Codex is a Fossil-native coding workbench for generating code, improving existing code, and improving its own ability to generate code. Its final purpose is simple:
Become an ever-improving coding tool.
A computer becomes useful for this when it has code and storage. Today Crystal is the representative tool for code, Fossil is the representative tool for storage and project memory, and a Linux host is the operating ground. These choices are pragmatic, not sacred — each should become more interchangeable when doing so improves the workbench without destroying present efficiency.
Coding and chat work is driven through a Fossil web UI, backed by a local Ollama model by default or a hosted provider, behind one consistent native-tool layer.
Why this exists
Human contributors and language models can both produce valuable work — and both produce work of uneven quality. Ollama-Codex should tolerate that reality without becoming vague, brittle, or unreviewable. Overall failure is treated as catastrophic, so the project follows three anti-failure axioms. They are standing invariants, never "done": every part and contribution is judged by them.
Clarity
Each part should become easier to name, understand, and judge. A part may have an imperfect or provisional purpose, but it should have a stated purpose; as understanding improves, that purpose can be clarified, subdivided, revised, or replaced.
Interchangeability
Each part should move, over time, toward being replaceable by something better. This does not mean abstracting everything immediately — it means avoiding unnecessary entanglement, exposing useful seams, and making replacement possible when it becomes worthwhile.
Accountability
Each part should be judged against its currently stated purpose. That purpose may later change, but while it stands it is binding. A contribution is not complete merely because it changes code; it must say what it is trying to accomplish and provide evidence that it does so.
No contribution without a claim; no claim without evidence.
Where to go next
This README is the terse source of truth for purpose and axioms. Everything else is referenced, not restated:
docs/architecture.md— definitions (Model / Provider / AgentRunner / Capability / Task vs Chat), the one-binary-many-interfaces shape, the seams, and interop boundaries. The single source of truth for how it fits together.docs/vision.md— the workbench model, the four causes, and the longer-range "why."CONTRIBUTING.md— contribution shape, evidence patterns, and practical rules (how Accountability is met in practice).docs/runtime-setup.md— runtime paths, deployment, and provider/API-key configuration.TODO.md— the one active milestone;BACKLOG.md— deferred ideas;PROGRESS.md— completed work.docs/review-checklist.md— recurring audits for axiom drift, accuracy, and gate depth.
Project state
Ollama-Codex is under active development. The current visible direction is to replace fragile coordination through external chat, CLI sessions, and tmux panes with a Fossil-hosted web UI that preserves context, coordinates model-backed tasks, shows progress, and supports review. The workbench should become useful before it becomes general.