Every feature built for AI-native development
Context Manager captures your organizational intelligence, compiles it into packages your AI tools understand, and keeps everything versioned and deterministic.
Context Hierarchy
Company > Product > Repo
Organize your organizational knowledge in a three-tier hierarchy. Company-level standards cascade to every product, product-level specs cascade to every repo. Bottom-up overrides let repos customize anything without breaking the chain.
Capabilities
- Three-level hierarchy: Company, Product, Repo
- Bottom-up override semantics (REPLACE, APPEND, PREPEND)
- Polymorphic context documents at every level
- Slug-based navigation for human-readable URLs
AI Wizard
AI-assisted context creation
Don't start from a blank page. The AI Wizard analyzes your existing documentation, meeting transcripts, and codebases to generate structured context documents. It understands your domain and asks the right follow-up questions.
Capabilities
- AI-generated context documents from raw inputs
- Company, product, feature, and codebase analysis wizards
- Iterative refinement through conversation
- Validates against existing context to prevent conflicts
Compilation Engine
Deterministic, cached, versioned
The compilation engine merges your entire context hierarchy into a single .context/ package per repo. Same inputs always produce the same output. Content-hash versioning means AI tools detect changes instantly.
Capabilities
- Deterministic: same inputs = same version hash
- Cached by hash — no duplicate compilations
- Outputs .context/ packages for AI coding tools
- fileMap JSON field — no individual file storage overhead
Stories, Recipes & Plans
Structured development artifacts
Break features into stories with acceptance criteria. Define recipes — reusable implementation patterns. Generate plans that sequence work by priority and dependency. Everything feeds back into the context loop.
Capabilities
- Stories with acceptance criteria and effort estimates
- Reusable recipes for common implementation patterns
- AI-generated implementation plans per story
- Dependency-aware sequencing and prioritization
Version Control Integration
Push compiled context to repos
Connect your GitHub repositories and push compiled context packages directly. Every compilation creates a versioned commit. Your AI tools always have the latest context without manual file management.
Capabilities
- GitHub integration via OAuth or personal access tokens
- Auto-push compiled .context/ packages to repos
- Versioned commits with context hash in message
- Branch-aware compilation for feature branches
Team Collaboration
Role-based access, audit trails
Teams are a permissions concern, not a context concern. Role-based access controls who can edit what. A complete audit trail tracks every change, every compilation, and every decision across your organization.
Capabilities
- Role-based access control (Owner, Admin, Member, Viewer)
- Full audit trail for every change and compilation
- Team-scoped views without polluting the context hierarchy
- Activity feeds and change notifications
Override Management
Fine-grained merge control
The default merge mode is REPLACE — repo-level context overrides product-level for the same key. But you can opt into APPEND or PREPEND when you need additive behavior. Full visibility into what overrides what.
Capabilities
- Three merge modes: REPLACE, APPEND, PREPEND
- Visual diff between hierarchy levels
- Override conflict detection and warnings
- Preview compiled output before pushing
How compilation works
Your context hierarchy gets compiled into a single package per repo. Deterministic, cached, and versioned.
Company Context
Standards, guidelines
Product Context
Specs, architecture
Repo Context
Codebase-specific
Compilation
Merge + hash
.context/ Package
AI-ready output
Ready to give your AI tools real context?
Start with a free assessment or try Context Manager today.