Statewave icon

Statewave

Statewave is an open-source memory runtime for AI agents and LLM apps, storing durable, structured context with provenance across sessions.

Statewave

Open-source memory runtime for AI agents

Statewave is an open-source memory runtime for AI agents and LLM applications. It is built to store durable, structured context with provenance so applications can recall prior events across sessions instead of starting from scratch each time.

The product is self-hosted on Postgres and is positioned as infrastructure for production AI systems. Its workflow records raw episodes, compiles them into typed memories, and returns ranked, token-bounded context that can be injected into prompts or agent loops.

Statewave also includes governance primitives such as sensitivity labels, audit receipts, deletion, and policy filtering. The site emphasizes local or self-hosted deployment, an admin console, and integration paths for MCP clients, SDKs, and connectors.

Core capabilities

Append-only episode capture

Statewave turns chats, tool calls, decisions, and workflow events into append-only episodes that preserve the original interaction history.

Typed memory compilation

Heuristic or LLM compilers derive typed memories such as profile facts, episode summaries, procedures, and artifact references, with confidence and validity metadata.

Context assembly for prompts

The runtime assembles ranked, token-bounded context bundles with deterministic retrieval, recency and similarity signals, and provenance back to source episodes.

Governance and audit artifacts

Governance features include sensitivity labels, policy filtering, tamper-evident audit receipts, per-entry supersession status, and deletion workflows.

Developer integration surface

The developer stack includes Python and TypeScript SDKs, a REST API, Docker Compose deployment, and an MCP server for compatible clients.

Connector-based ingestion

Connectors normalize data from tools such as GitHub, Slack, docs, Zendesk, n8n, Zapier, and Jira into the same episode model.

Practical use cases

  • Persistent agent memory

    Build agents that remember prior decisions, support threads, or workflow history and return that context on the next turn instead of relying only on recent chat messages.

  • Team and project knowledge

    Pull issues, pull requests, docs, Slack threads, tickets, and workflow runs into a shared memory layer so teams can ground answers in project history and decisions.

  • Governed context for production systems

    Use provenance, sensitivity labels, and audit receipts when you need reviewable context assembly and controlled access to memories in production workflows.

  • Developer integration and local setup

    Connect the runtime to MCP-compatible clients, SDK-based applications, or self-hosted deployments when you want a local memory layer that fits into an existing developer workflow.

  • Support and operations memory

    Normalize external systems through connectors so support or ops workflows can recall customer histories, ticket status, and prior attempts across tools.

Pros and Cons

Pros

  • Open-source and Apache-2.0 licensed, which makes the runtime source-visible and self-hostable.
  • Runs on Postgres and can be started locally with a one-command setup that includes the API, admin console, and database.
  • Supports structured memory with provenance, so individual memories trace back to source episodes.
  • Includes governance-oriented features such as sensitivity labels, audit receipts, and deletion workflows.
  • Offers multiple developer entry points, including SDKs, REST API access, an MCP server, and modular connectors.

Cons

  • The pricing page currently returns a 404, so the public site does not provide a clear pricing model or plan structure.
  • Some capability details are described on product and connector pages, but the site still leaves parts of the implementation and supported ecosystem to the docs.

FAQ

Can I run Statewave locally?

Statewave is designed to run locally or in a self-hosted deployment. The developer flow shows a one-command setup that starts the API, admin console, and Postgres, and it can be wired into MCP clients without creating an account.

What does Statewave output to my app?

Statewave stores raw episodes, compiles typed memories, then retrieves ranked, token-bounded context for the model. The product page describes a deterministic loop of record, compile, context, and govern.

How do developers integrate it?

The site highlights Python and TypeScript SDKs, a REST API, Docker Compose deployment, OpenTelemetry, and an MCP server for clients such as Claude Code, Claude Desktop, Cursor, VS Code Copilot, and Codex CLI.

What kinds of data can I feed into it?

The connectors page lists sources such as GitHub, Slack, Discord, docs and Markdown, Zendesk, Intercom, Freshdesk, Gmail, Jira, n8n, Zapier, and database records. Each connector normalizes incoming events into Statewave's episode model.

Is pricing available on the site?

The website does not show pricing details. A pricing URL currently returns a 404 page, so the source material only confirms the product is open source and self-hosted, not how any paid offering is structured.

Quick Facts

Category
Developer Tool / AI Infrastructure
Primary use
Memory runtime for AI agents and LLM apps
Deployment
Self-hosted on Postgres; local setup supported
License
Apache-2.0
Developer interfaces
Python SDK, TypeScript SDK, REST API, MCP server
Source domain
statewave.ai