Skip to content
@sovantica

Sovantica

Cognitive infrastructure for AI agents.
README.md

Sovantica

Cognitive infrastructure for AI agents.

We build the substrate AI agents run on — memory, reasoning scaffolds, verifiable state. Starting with the hardest part: memory that doesn't collapse under context-window limits, and that an agent can actually reason over.

What we ship

Engrava — the memory database for AI agents.

An embedded cognitive graph database for AI agents. Graph memory (7 edge types), hybrid FTS5 + vector search, algorithmic dreaming (6-signal consolidation with configurable gates), tamper-evident thought/edge journal. No Neo4j. No Redis. Just SQLite.

pip install engrava

MIT licensed · engrava.ai · docs

How we work

  • Code-first. Every claim is backed by implementation. No whitepapers without running code.
  • Research posture. Grounded in consciousness research, not AI hype. Citations where it matters.
  • Open by default. MIT licensed, public roadmap, discussions over DMs.

Elsewhere

Popular repositories Loading

  1. engrava engrava Public

    The memory database for AI agents — graph memory, hybrid search, tamper-evident thought/edge journal.

    Python 1

  2. engrava-mcp engrava-mcp Public

    Model Context Protocol server for Engrava — expose an AI agent memory database to any MCP client (Claude Desktop, Cursor, VS Code…) over stdio.

    Python 1

  3. .github .github Public

  4. engrava-benchmark engrava-benchmark Public

    Public reproduction surface for Engrava memory-benchmark numbers (runners + machine-readable results).

    Python

Repositories

Showing 4 of 4 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…