v0.2.0 — Phase 1 close

Public install + docs surface for AgentBay. Phase 1 ships:

• per-agent install recipes (Claude Code, Codex CLI, Cursor, OpenClaw, Hermes)
• competitor comparison pages (vs. GBrain, Mem0, Letta, Zep)
• public scorecards (LOCOMO + AgentBay-native rich-prose, three-row ablations) mirrored from the private app
• runnable first-recall examples (Python, TypeScript, MCP JSON)
• README polished around the GBrain wedge — one user, one brain vs. teams, projects, governance — with the honest "Where we lose" callout

GBrain is one user, one brain. AgentBay is teams, projects, governance.

Star this repo to follow the public install surface, and sign in at https://aiagentsbay.com when you are ready to sync agent memory with a team.

v0.1.0 — public launch

First public release of the AgentBay project page — install docs, comparison pages, scorecards, and MCP recipes for every agent host.

GBrain is one user, one brain. AgentBay is teams, projects, governance.

What's in this repo

• README with the wedge positioning, three-line install (Python / TypeScript / MCP), and the in-progress comparison table vs GBrain, Mem0, Letta, Zep.
• Cross-references to the SDK source repos:
  • Python SDK source — thomasjumper/agentbay-python
  • TypeScript SDK — agentbay on npm (source repo coming)
  • MCP server — aiagentsbay-mcp on npm (source repo coming)
• Hosted dashboard — aiagentsbay.com.
• Issue templates and a contributing guide scoped to docs / install recipes / comparison content.

What's coming next

The following directories are scaffolded but not yet populated — these are the next public PRs:

• + 'docs/install/' + — per-agent install recipes (Claude Code, OpenAI Codex CLI, Cursor, OpenClaw, Hermes).
• + 'docs/scorecards/' + — public mirror of our retrieval evaluation scorecards. Methodology and reproduction instructions included.
• + 'docs/vs/' + — comparison pages vs GBrain, Mem0, Letta, Zep with honest concessions and concrete capability claims.
• + 'examples/' + — runnable code samples for the Python SDK, TypeScript SDK, and MCP server.

Install

```bash
pip install agentbay # Python SDK (PyPI 1.7.1)
npm install agentbay # TypeScript SDK
npx aiagentsbay-mcp # MCP server for Claude Code, Cursor, Codex CLI, Windsurf, ChatGPT
```

Then sign in at aiagentsbay.com for teams, projects, and shared memory across machines.

Project structure

• This repo (public): install + docs + marketing surface. MIT.
• agentbay-python (public): Python SDK source. MIT.
• agentbay-app (private): hosted Next.js application source.