Bezalel
One MCP endpoint gives every AI agent you run, across Claude Code, Cursor, or Codex CLI, the same shared memory, email inbox, and connectors, so switching frameworks stops meaning starting over.
The verdict on Bezalel: Sharing memory, email, and connector access across multiple AI agent frameworks without re-wiring each one Bezalel is squarely a developer tool, if you're not already running agents through Claude Code, Cursor, or a similar MCP-compatible runtime, this isn't for you. Pricing: Free sign-up during alpha (pricing not yet published) / self-hostable via Docker for teams that want to run their own plane. Last reviewed: August 2026.
Best For
Sharing memory, email, and connector access across multiple AI agent frameworks without re-wiring each one
Standout Feature
Agents keep the same memory, inbox, and connections no matter which MCP-compatible runtime they're running in
Verdict
A clever piece of infrastructure for developers running personal agents across multiple tools, not something a non-technical user will set up alone.
Alternatives
Overview
Bezalel solves a specific annoyance for anyone running personal AI agents across more than one framework: every time you switch from Claude Code to Cursor or Codex CLI, you re-wire the same memory, email, and connector setup from scratch. Bezalel inverts that by hosting your agent's capabilities, long-term memory, a real email inbox, a spend ledger tied to bank data, iMessage texting, a cloud desktop, disposable code sandboxes, and hundreds of third-party app connectors, behind one MCP endpoint with a single URL and bearer token. Point any MCP-compatible agent runtime (Claude Code, Codex CLI, Cursor, OpenClaw, and others) at that endpoint, and it inherits the same memory, inbox, and connections your other agents already built up, so switching frameworks becomes a config change instead of a rebuild.
Inbound events matter here too: a text message or an email can wake an agent instead of it having to poll for updates. The service is self-hostable as a single Docker container if you'd rather keep everything on your own infrastructure. Bezalel is in alpha, sign-up is free while pricing is still unpublished, and features like virtual card controls are listed as coming soon.
Our Take
Bezalel is squarely a developer tool, if you're not already running agents through Claude Code, Cursor, or a similar MCP-compatible runtime, this isn't for you. For the audience it is for, the pitch is legitimately useful: instead of re-teaching every new agent framework your email setup and connector list, you point it at one endpoint and it already knows. It's alpha software from a small team, some features like virtual cards are still coming soon, and giving any single service standing access to your email, bank data, and texting is worth weighing carefully regardless of how convenient the setup is.
Key Features
- Single MCP endpoint for all agent capabilities
- Shared long-term memory across agent frameworks
- Real email inbox with inbound event triggers
- iMessage integration with agent-waking texts
- Cloud desktop and disposable code sandboxes
- Self-hostable via a single Docker container
- • Switching agent frameworks no longer means rebuilding memory and connectors
- • Inbound events (texts, emails) can wake an agent instead of polling
- • Self-hostable option for teams that want to keep data off Bezalel's infrastructure
- • Built for developers running MCP-compatible agents, not a consumer tool
- • Alpha stage with some advertised features (like virtual cards) not yet live
- • Requires connecting sensitive access (email, bank data, texting) to a small, new company
People Also Use
Other Developer Tools tools builders reach for alongside Bezalel.
Hugging Face
Host, share, and download open models, datasets, and demo apps, model discovery and deployment in a few clicks instead of a research project.
Dialogflow
Build rule-based or generative conversational agents for chat and voice that plug into Google Cloud's NLU and generative AI stack, billed per request or session. Increasingly marketed as Conversational Agents.
LangChain
Assemble LLM-powered apps and agents from composable building blocks, with LangSmith adding tracing, evaluation, and deployment. Platform rebranded. LangGraph Platform is now LangSmith Deployment.
Ollama
Run open-weight language models directly on your own machine with a single command, or shift to hosted GPUs via Ollama Cloud when local hardware isn't enough.
Supabase
Get a Postgres database, auth, storage, and edge functions in one backend, with an AI Assistant and MCP integrations, so small teams ship apps without managing infrastructure.
LlamaIndex
Connect your own data to LLMs for retrieval-augmented generation. LlamaParse (formerly LlamaCloud) automates document parsing, extraction, and indexing for agentic workflows.