OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Honcho

Honcho is an open-source Python memory library and FastAPI server for building stateful agents. It stores conversations, events, documents, and tool traces, processes them in the background, and exposes peer representations, session context, search results, and natural-language answers through SDKs, a CLI, or server endpoints.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Honcho — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Honcho when an agent must carry information across interactions and retrieve processed context rather than relying only on the current prompt or basic chunk matching.

Who it is for

  • Developers building stateful conversational agents
  • Teams adding persistent memory to AI products
  • Builders of agentic RAG and context-engineering systems
  • Developers integrating memory through Python, TypeScript, or CLI tools

Documented uses

  • Store conversations, events, documents, and tool traces for agent use.
  • Maintain changing representations of users, agents, groups, projects, and ideas.
  • Retrieve session context, search results, or peer-level insights for an LLM call.
  • Expose agent memory through a managed service, local stack, or self-hosted server.
Before you choose

Requirements & limits

The repository documents a managed service at api.honcho.dev, a local CLI-started stack, and self-hosting. The managed quickstart uses an API key and workspace ID; the local example points the SDK to http://localhost:8000.

  • Choose the Python SDK for the documented Python integration, or the TypeScript SDK for npm-based applications.
  • Use the CLI when running a local stack or inspecting a deployment with the documented workspace commands.
  • The service separates asynchronous background reasoning from storage and querying, so applications need to account for that processing model.
  • The AGPL-3.0 license is a material consideration for teams incorporating the public repository.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Add stored context to an agent with Honcho

Set up Honcho, store messages on a session, retrieve context, and pass that context to an LLM.

Before you start

  • Python
  • The honcho-ai package
  • A Honcho API key and workspace ID for the managed service, or a self-hosted Honcho URL
  • An OpenAI client and model for the documented example

Steps

  1. Install the honcho-ai package with pip, uv, or Poetry.

  2. Create a Honcho client with a workspace ID and API key, or configure it with a self-hosted base URL.

  3. Create peers and a session, then add peer messages to the session.

  4. Query peer knowledge or session context and convert the context for an LLM call.

Expected outcome

The documented flow supplies an LLM call with context retrieved from stored Honcho session activity.

Guide limitations

  • The example's background reasoning stage is asynchronous.
  • The documented LLM example uses the OpenAI client; other model or framework integrations are described but not specified in this recipe.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
2 October 2026
Original website title
plastic-labs/honcho · Official homepage title · retained in its source language
How we review sourcesSuggest a correction