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

Hivemind

An Apache-2.0 TypeScript library and project for shared agent memory, trace retrieval, and reusable skills.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Hivemind — 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 it when coding agents need a common memory layer or when successful traces should become reusable skills across a team.

Who it is for

  • Teams using multiple AI coding agents
  • Developers who want prior agent work available in later sessions
  • Builders creating shared or cross-agent memory workflows

Documented uses

  • Capture agent prompts, tool calls, and responses as structured traces
  • Search previous agent experiences with lexical and semantic retrieval
  • Turn repeated patterns into reusable SKILL.md files
  • Propagate skills and memory across supported agents and machines
Before you choose

Requirements & limits

The repository is public and licensed under Apache-2.0.

  • Which supported assistant needs integration: Claude Code, OpenClaw, Codex, Cursor, Hermes Agent, pi, or Claude Cowork Alpha
  • Whether the workflow requires lexical and semantic retrieval or can use the documented lexical fallback when embeddings are off
  • Whether storage should use the documented GCS, Azure, S3, or on-premises bring-your-own-cloud approach
  • Whether headless installation is needed through a token rather than browser sign-in
  • Whether the reported LoCoMo benchmark configuration is relevant to the intended workload

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

Install Hivemind and wire a supported assistant

Set up Hivemind hooks for supported agents and check the resulting configuration.

Before you start

  • macOS or Linux, Windows PowerShell, or an environment with Node 22+ and a writable npm prefix for the npm route.
  • A browser sign-in flow or an account token for enabling shared memory.

Steps

  1. On macOS or Linux, run the documented shell installer; on Windows, run the documented PowerShell installer.

  2. Alternatively install the package globally with npm and run hivemind install.

  3. Use the assistant-specific install command when configuring only Claude Code, Codex, OpenClaw, Cursor, Hermes, pi, or Claude Cowork Alpha.

  4. Restart the assistants after installation, then run hivemind status to check what is wired up.

  5. For headless installation, provide a token through HIVEMIND_TOKEN or --token; if no token is supplied in a non-interactive shell, run hivemind login later.

Expected outcome

The installer detects or configures the selected assistant integrations and reports their wiring status.

Guide limitations

  • The npm route requires Node 22+ and a writable npm prefix.
  • The supplied material documents setup and status checking, but not a complete task-execution workflow after installation.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
30 September 2026
Original website title
activeloopai/hivemind · Official homepage title · retained in its source language
How we review sourcesSuggest a correction