
Knownbase
Provides persistent, searchable project memory for AI coding agents to store and retrieve architecture decisions, debugging discoveries, rejected approaches, constraints and handoffs
The product is a persistent, searchable project memory layer for AI coding agents that stores architecture decisions, debugging discoveries, rejected approaches, constraints, conventions and handoffs so knowledge survives session compaction and tool switches. It is for software developers and engineering teams in businesses that use Claude Code, Codex, Cursor and other MCP-compatible agents. It is positioned against instruction files like CLAUDE.md, offering versioned, project-scoped notes that agents search and retrieve on demand instead of reloading a growing file, shared across agents and teammates via MCP.
Key features
- Persistent project memory for agents
- Search and filtering by text and metadata
- Projects with status workflows and tags
- Revision history with restore
- Backlinks and note links
- Shared workspaces with roles
- In-thread comments on notes
- Read-only share links
- Live updates via server-sent events
- Outbound webhooks on note changes
- Templates for decisions and handoffs
- Workspace isolation per tenant
- Two-factor auth with TOTP
- Scoped workspace API keys
- No social media activity within the last 30 days
- Blog
- Marketplace
- API
- Docs
- Software developers
- Engineering teams
- DevOps sre teams