User documentation

Everything you need to install, operate, and get value from ContextMint — Engine on your machine, the VS Code extension, and the human-in-the-loop workflow. Start with getting started if you are new; use these chapters for depth.

Installation & setup

Prerequisites, local-first path, verify steps — plus installation support (first-line fixes).

Chat, modes & context lanes

Ask / Plan / Agent, Repo / Work / Hybrid lanes, attachments, sessions, and auto-suggest.

Image evidence & ARGUS

Paste screenshots, local Ollama vision, Context Lens preview, and structured visual audit.

Context Lens & packs

Gate 1 preview, relevance scores, @pack activation, scan & learn, stale packs.

Engine operator guide

Tray app tabs: Overview, Server, Ollama, Models, Indexing, Quality, Knowledge, Logs, Settings.

Indexing & readiness

Progressive RAG while indexing, readiness phases, priority files, hot paths.

Patches & governance

Assisted actions, patch preview (gate 2), quality map, DRY findings, audit trail.

Knowledge & governance

Engine Knowledge tab — structural registries, activity audit, persistence, worked examples.

Settings reference

VS Code contextmint.* (149 keys) and API server defaults (1025 keys) with usage guides.

Context budget calculator

Scale server max_context_tokens and see overview pack, seed, and retrieval slot knobs update proportionally.

BYOK & enterprise

Cloud API keys, redaction, shared on-prem server, OIDC, remote SSH path alignment.

Sovereign server

Platform deploy with OpenAI-compatible inference (vLLM), OIDC, and path-aligned workspaces.

Glossary

v1.0 terms plus Verification Engine jargon: blast radius, MCP, drift, decision graph, and more.

Troubleshooting

Symptom → cause → fix for every common problem across install, chat, and indexing.