Skip to content
Home

Sekha

Provides persistent, searchable memory for AI assistants, with structured metadata, semantic tagging, and a 4-phase assembly pipeline for context retrieval.

Sekha is a context broker that provides persistent memory for AI assistants, solving the problem of context loss in chat threads. It targets developers and engineering teams in B2B settings, offering a self-governed memory system that agents consult but do not own. Unlike platform-bound memory solutions, it is delivered as an open-source, self-hosted system with REST endpoints and MCP tools, ensuring user ownership and control.

Key features

  • Library cards with semantic tagging
  • SQLite storage you own and query
  • Native folder/label system
  • 4-phase assembly pipeline
  • Citation metadata injection
  • REST endpoints and MCP tools
  • Importance scoring and recency decay
  • Semantic search across conversations
  • Pinned conversations and preferred labels
  • Token budget packing
  • No social media activity within the last 30 days
GTM channels
  • Community
  • API
  • Docs
ICP
  • Software developers
  • Engineering teams
  • Startups