Skip to content
MCP server

Keyblind

Encrypted local secrets vault that lets AI agents use keys without seeing them

Runs locallyStatus and history
Install
Pricing
Freemium
Platforms
CLI, macOS, Linux
Runs
Locally
Version
0.8.8
Listing updated
2 days ago

About Keyblind

Keyblind is a freemium MCP server for security. It runs locally on your machine and is started by your AI client.

The idea is that an agent asks for a secret by name and the plaintext value never lands in the chat transcript. Secrets sit in an encrypted SQLite vault (AES-256-GCM, with a key bound to the machine identity), and a sandbox mode swaps the values in a .env file for deterministic fake ones so AI code review sees clean diffs.

Beyond storage, the project bundles a lot: TOTP and HOTP code generation, expiring and view-limited share links, a dead man's switch that releases the vault to contacts if you stop checking in, git-safe team vaults, and Slack or Discord alerts for expiry. Seven backends are supported: local, 1Password, Bitwarden, environment variables, AWS, GCP and Azure.

A keyblind quickstart command sets up the vault and configures Claude Code, Cursor, Copilot, Windsurf, Cline or Zed. The registry entry lists 27 tools, and an optional Touch ID gate needs a Pro or Team licence. The code is MIT licensed.

Install

Run in a terminal

claude mcp add keyblind -- npx -y keyblind

Add to claude_desktop_config.json

{
  "mcpServers": {
    "keyblind": {
      "command": "npx",
      "args": [
        "-y",
        "keyblind"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "keyblind": {
      "command": "npx",
      "args": [
        "-y",
        "keyblind"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "keyblind": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "keyblind"
      ]
    }
  }
}

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

{
  "mcpServers": {
    "keyblind": {
      "command": "npx",
      "args": [
        "-y",
        "keyblind"
      ]
    }
  }
}

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

{
  "mcpServers": {
    "keyblind": {
      "command": "npx",
      "args": [
        "-y",
        "keyblind"
      ]
    }
  }
}

Tools this server provides (17)

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

resolve_secret
Resolves a secret at runtime with the value hidden from the transcript.
store_secret
Encrypts and stores a secret.
list_secrets
Lists secret names without values.
delete_secret
Deletes a secret.
sandbox_env
Replaces .env values with deterministic fakes.
unsandbox_env
Restores real .env values from the vault.
audit_log
Shows the secret resolution audit trail.
totp_code
Generates a TOTP code for a stored config.
totp_store
Stores a TOTP configuration from an otpauth URI.
totp_list
Lists stored TOTP configurations.
totp_delete
Deletes a TOTP configuration.
create_share_link
Creates an encrypted, expiring share link for a secret.
receive_share
Receives and decrypts a shared secret.
deadman_status
Checks the dead man's switch status.
deadman_checkin
Resets the dead man's switch timer.
team_init
Creates a shared team vault.
team_push
Pushes a local secret to the team vault.

What it does

Runtime secret resolution
Agents reference secrets by name; values stay out of the transcript.
Sandboxed .env files
Real values are replaced by deterministic fakes and restored on demand.
Seven backends
Local vault, 1Password, Bitwarden, env vars, AWS, GCP, Azure.
Built-in 2FA codes
Stores TOTP configs and generates codes.
Share links and dead man's switch
Time-limited shares and automatic hand-off if you disappear.
Team vaults
Passphrase-protected shared vaults, with SSO/OIDC on the Team plan.

Good for

  • Stop an AI coding assistant from reading real keys in .env
  • Run commands with secrets injected as environment variables
  • Share a one-off credential with a teammate through an expiring link
  • Sandbox .env files before AI code review in CI
  • Generate 2FA codes for a stored account from an agent

Pricing

Freemium

Free for 5 secrets; Pro $79/year; Team plan adds SSO

Prices change. Check keyblind.dev before you buy.

Keyblind FAQ

What does it cost?

The site lists a free tier limited to 5 secrets, and Pro at $79 a year for unlimited secrets; Team adds SSO and role-based access.

Which editors does it support?

Claude Code, Cursor, Copilot, Windsurf, Cline and Zed.

Where are secrets stored?

In a local encrypted SQLite vault by default, with optional backends such as 1Password or AWS.

How is it licensed?

The README comparison table states MIT.

Is Keyblind free?

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

How do I add Keyblind to Claude Code?

Run this command in your terminal: claude mcp add keyblind -- npx -y keyblind

What are the best alternatives to Keyblind?

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

Alternatives to Keyblind

  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