CueAPI MCP Server for Claude
Official CueAPI stdio MCP server for scheduling agent work, claiming executions, and reporting evidence-backed outcomes with verified handoffs across time and environments.
Open the source and read safety notes before installing.
Safety notes
- cueapi_delete_cue and cueapi_fire_cue are destructive or side-effecting; confirm targets before approval.
- Worker claim tools mutate execution state; misconfigured heartbeats can lose leases on in-flight jobs.
- Webhook payloads from fired cues may trigger external systems; validate callback URLs.
- API keys prefixed cue_sk_* grant account-level scheduling and execution access.
Privacy notes
- Execution history, callback URLs, and outcome evidence may include internal system identifiers.
- Store CUEAPI_API_KEY in MCP env vars, not in chat, issues, or committed config files.
- Self-hosted deployments keep data within your infrastructure per your CueAPI core configuration.
Prerequisites
- CueAPI account and API key prefixed with cue_sk_ from https://cueapi.ai.
- Node.js 18+ with npx available for @cueapi/mcp.
- Claude Desktop, Claude Code, Cursor, Zed, or another MCP host with stdio support.
- Optional CUEAPI_BASE_URL when self-hosting CueAPI core instead of the hosted API.
Schema details
- Install type
- cli
- Reading time
- 4 min
- Difficulty score
- 40
- Troubleshooting
- Yes
- Breaking changes
- No
- Scope
- Source repo
- Estimated setup
- 10 minutes
- Difficulty
- intermediate
- Website
- https://cueapi.ai
Full copyable content
{
"mcpServers": {
"cueapi": {
"command": "npx",
"args": ["-y", "@cueapi/mcp"],
"env": {
"CUEAPI_API_KEY": "cue_sk_YOUR_KEY"
},
"type": "stdio"
}
}
}About this resource
Overview
CueAPI is an open coordination layer for multi-step agent work. The official cueapi/cueapi-mcp package exposes CueAPI as a stdio Model Context Protocol server so hosts like Claude Desktop, Cursor, and Zed can schedule cues, inspect executions, claim worker jobs, and report evidence-backed outcomes from chat.
CueAPI closes handoffs with structured proof—external IDs, result URLs, or artifacts—so silent failures are easier to detect across long-running workflows.
Features
- Create recurring cron cues or one-time
atschedules. - List, pause, resume, update, fire, and delete cues.
- List and fetch execution history with state and outcomes.
- Claim executions individually or claim-next with optional task filters.
- Send execution heartbeats to extend worker leases.
- Report write-once outcomes with verifiable evidence.
Use Cases
- Schedule a daily digest webhook without maintaining a separate cron service.
- Claim the next available worker execution from an MCP-hosted agent session.
- Fire an ad-hoc cue with a payload override for a one-shot integration test.
- Inspect the last five runs of a production cue before changing its schedule.
- Report an outcome URL after an external job completes.
Installation
npm global install
npm install -g @cueapi/mcp
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"cueapi": {
"command": "npx",
"args": ["-y", "@cueapi/mcp"],
"env": {
"CUEAPI_API_KEY": "cue_sk_..."
}
}
}
}
Claude Code
claude mcp add cueapi --env CUEAPI_API_KEY=cue_sk_YOUR_KEY -- npx -y @cueapi/mcp
Restart the host after updating MCP configuration.
Configuration
{
"mcpServers": {
"cueapi": {
"command": "npx",
"args": ["-y", "@cueapi/mcp"],
"env": {
"CUEAPI_API_KEY": "cue_sk_YOUR_KEY",
"CUEAPI_BASE_URL": "https://api.cueapi.ai"
},
"type": "stdio"
}
}
}
Set CUEAPI_BASE_URL only when pointing at a self-hosted CueAPI core deployment.
Examples
Schedule a daily cue
Create a CueAPI cron cue that fires every weekday at 09:00 UTC and posts to my webhook.
Inspect executions
List the last five executions for cue_abc123 and summarize failures.
Claim and report
Claim the next available execution for task_name=data-sync, process it, then report the outcome URL.
Security
- Treat
cue_sk_*keys like production credentials; rotate them if exposed. - Confirm callback URLs before firing cues that dispatch to external webhooks.
- Use heartbeats during long-running claims so another worker cannot steal the lease unexpectedly.
Troubleshooting
Invalid API key
Generate a fresh cue_sk_* key at cueapi.ai and update the MCP env block.
404 on pause or resume older docs
Current tools use PATCH /v1/cues/{id} with status paused or active; upgrade @cueapi/mcp if you run an old release.
Claim starvation
Prefer cueapi_list_claimable_executions with server-side task or agent filters instead of client-side filtering.
Self-hosted connection errors
Set CUEAPI_BASE_URL to your core API origin and verify TLS or VPN reachability from the MCP host.
Source citations
Add this badge to your README
How it compares
CueAPI MCP Server for Claude side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.
| Field | CueAPI MCP Server for Claude Official CueAPI stdio MCP server for scheduling agent work, claiming executions, and reporting evidence-backed outcomes with verified handoffs across time and environments. Open dossier | MCP Registry MCP Server MCP server that lets Claude and other MCP clients search and page through the official MCP Registry from a local stdio server. Open dossier | Adeu MCP Server for Claude Adeu MCP server and SDKs translate DOCX files to token-efficient Markdown with CriticMarkup, apply safe Track Changes edits, and optionally integrate with live Microsoft Word on Windows. Open dossier | AgentTrust MCP Server for Claude AgentTrust stdio MCP server giving AI agents verified identity with email, instant messaging, and cloud file storage across 19 Ed25519-signed tools. Open dossier |
|---|---|---|---|---|
| Trust | ||||
| Install risk | Review first | Review first | Review first | Review first |
| Notes | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ | Safety ✓ Privacy ✓ |
| Category | mcp | mcp | mcp | mcp |
| Source | source-backed | source-backed | source-backed | source-backed |
| Author | CueAPI | Jun Han | Adeu | AgentTrust |
| Added | 2026-06-14 | 2026-06-06 | 2026-06-14 | 2026-06-14 |
| Platforms | Claude CodeCodexCursorClaude Desktop | Claude CodeClaude Desktop | Claude CodeClaude Desktop | Claude CodeClaude Desktop |
| Source repo | — | — | — | — |
| Safety notes | ✓cueapi_delete_cue and cueapi_fire_cue are destructive or side-effecting; confirm targets before approval. Worker claim tools mutate execution state; misconfigured heartbeats can lose leases on in-flight jobs. Webhook payloads from fired cues may trigger external systems; validate callback URLs. API keys prefixed cue_sk_* grant account-level scheduling and execution access. | ✓Registry search results can point to arbitrary third-party MCP servers, packages, remotes, and repositories; review each result before installing anything it recommends. The server shells through `npx`, so pin a package version if you need reproducible installs instead of `@latest`. A custom `registry_url` can point the server at a private registry; only use trusted registry endpoints. | ✓`process_document_batch` can modify DOCX content and apply Track Changes when edits pass validation. `finalize_document` can strip metadata and lock documents; review output before external distribution. Live Word integration edits the active document in memory; save backups before batch operations. Adeu Cloud email workflows fetch external attachments; treat counterparty documents as untrusted input. | ✓Email tools can send outbound mail from your `@agenttrust.ai` address. Messaging tools can contact other agents and escalate to humans via HITL flows. Drive tools upload, download, and delete files—confirm destructive operations. Ed25519 signing keys are generated locally; back up `~/.agenttrust` before rotation. |
| Privacy notes | ✓Execution history, callback URLs, and outcome evidence may include internal system identifiers. Store CUEAPI_API_KEY in MCP env vars, not in chat, issues, or committed config files. Self-hosted deployments keep data within your infrastructure per your CueAPI core configuration. | ✓Search terms, pagination cursors, custom registry URLs, registry responses, package names, repository URLs, and server metadata may be visible to the MCP client and model provider. Private registry URLs or internal server metadata can be exposed in prompts, logs, transcripts, and client config if used directly. | ✓Local Adeu tools process documents on your machine and do not phone home with file contents by default. Document text sent to your LLM provider is subject to that provider's data handling policies. Optional Adeu Cloud routes large validation or email workflows through Adeu servers under separate terms. | ✓Email bodies, attachments, messages, and uploaded files are processed by AgentTrust. Signing keys and API keys are stored locally with 0600 permissions but still require host protection. Outbound email from address is enforced server-side to your agent's `@agenttrust.ai` identity. |
| Prerequisites |
|
|
|
|
| Install | | | | |
| Config | | | | |
| Citations | ||||
| Claim | Unclaimed | Unclaimed | Unclaimed | Unclaimed |
Signals
Loading live community signals…
A short, calm digest of reviewed Claude resources. Unsubscribe any time.