Skip to content
MCP server

qbash

Build, schedule and run Python AI tasks in qbash and share memory across agents

Install
Pricing
Freemium
Platforms
API
Runs
Remotely
Version
1.0.0
Listing updated
2 days ago

About qbash

qbash is a freemium MCP server for developer tools. It runs as a hosted remote server, so there is nothing to install locally.

qbash's server connects an agent to a platform where automations are Python tasks. You can hand over written requirements, or a SPEC.md with tests, and the task builder writes the task with inline prompts where AI judgement is needed. Tasks can run on a schedule or behind a webhook, and every save is a version you can roll back to.

The server can also run tasks and read output and step-by-step logs, with a dry run option that holds back connector writes, memory writes and non-GET web requests while reads and AI prompts still run and cost money. Connector calls such as Reddit, HubSpot or Slack use accounts you connected, and the first use of each account needs your approval for 1 hour up to Always.

Shared memory folders let Claude Code, Codex and other agents read and write the same documents. Edits are versioned, and folders can require a person to accept a review. Custom servers can expose a chosen set of tasks, accounts and folders to a team.

Install

Run in a terminal

claude mcp add --transport http qbash https://mcp.qbash.com/mcp

Add to Settings → Connectors → Add custom connector

https://mcp.qbash.com/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "qbash": {
      "url": "https://mcp.qbash.com/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "qbash": {
      "type": "http",
      "url": "https://mcp.qbash.com/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "qbash": {
      "serverUrl": "https://mcp.qbash.com/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "qbash": {
      "type": "http",
      "url": "https://mcp.qbash.com/mcp"
    }
  }
}

Tools this server provides (12)

Taken from the project documentation; the running server may differ.

list_tasks
List tasks you can run with their latest run status
run_task
Start a production or dry run of a task
build_task
Build a new task with the task builder
check_build
Read progress of a task build
get_task_result
Read a run's result and held-back writes
get_task_log
Read a run's logs step by step
read_task
Read a task before reusing or revising it
call_connector_read
Read through a connected account
call_connector_write
Write through a connected account
edit_memories
Edit memory documents by id
memory_reviews
List open memory reviews
request_connector
Get a link to connect an account

What it does

Task builder
Turns requirements or a SPEC.md into a Python task, versioned on each save.
Run and inspect
Run deployed tasks and read results and logs step by step.
Dry runs
Hold back writes to see what a task would send.
Connector access
Call tools like Reddit, HubSpot or Slack through connected accounts with timed approval.
Shared memories
Agents read and write common memory folders with versioned edits and optional review.
Custom servers
Expose a fixed set of tasks, accounts and folders to a team.

Good for

  • Have an agent write and deploy a scheduled Python task
  • Test a task safely before it posts or updates anything
  • Share notes between Claude Code and Codex sessions
  • Give a team one tool that wraps a connector account
  • Review agent-proposed memory edits before they go live

Pricing

Freemium

Free: 100 runs a month and 3 tasks; Developer and Team plans, prices not shown

Prices change. Check qbash.com before you buy.

qbash FAQ

How do I sign in?

With your qbash account via OAuth, or with an access token sent as a Bearer header.

Can I try it free?

Yes, the Free plan includes 100 runs a month and 3 tasks.

Who pays for model usage?

Models use your own provider keys; model and third-party API charges are separate.

How long does connector approval last?

You choose 1 hour, 24 hours, 7 days, 30 days or Always.

Is qbash free?

qbash has a free plan, with paid plans for more usage or features.

How do I add qbash to Claude Code?

Run this command in your terminal: claude mcp add --transport http qbash https://mcp.qbash.com/mcp

What are the best alternatives to qbash?

The closest alternatives to qbash are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to qbash

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools