Arcade MCP — Watchlist

MCP100 Assessment

Watchlist / Tracking

Why we’re tracking it

Follow hosted tool deployment and provider integrations as the platform expands

Reviewed Oct 5, 2026 · Documentation review; complete setup workflow not independently tested

Best fit

Building custom authorized MCP servers and tool-evaluation workflows

What we know

Python framework for custom MCP tools, OAuth-aware authorization, evaluation and Arcade Cloud deployment.

Maintenance / activity evidence

Observed Oct 2, 2026 (UTC)fix(arcade-core): describe every TypedDict field in tool schemas ( 950)

Tracking evidence observed Oct 5, 2026. These dated signals do not establish current maintenance. Reviewed enrichment confidence: high.

Activity source · commits ↗Activity source · releases ↗

Verification gaps / watch-outs

Installation and comprehensive security have not been independently audited.
Tool-declared upstream scopes. Writes: Depends on authored tools.
OAuth/API keys/resource-server auth.
Sensitive access: Upstream service data/credentials

Framework risk is dominated by custom tools, requested scopes and deployment mode.

Security source · repository notice ↗Security source · policy file ↗

Full authority, access and evidence limits are available in Review context and Access & risk.

Popularity

1,046 repository stars

Stars observed Oct 6, 2026.

30-day package downloads intentionally left null because the package-stats provider ran out of credits during Batch 3 QA.

All limitations / Review context

Which hosted MCP providers and authorization paths are currently documented?

The bounded assessment reviewed the official README, repository notices, security policy path, and security page for known material public-tracking blockers. No such blocker was identified. This is a public-tracking assessment, not an installation safety certification.

Framework risk is dominated by custom tools, requested scopes and deployment mode.
Tool-declared upstream scopes.
OAuth/API keys/resource-server auth.
Hosted secrets are injected at runtime; custom tools decide data flow.
Reviewed write actions: Depends on authored tools
Sensitive access: Upstream service data/credentials
Installation, client compatibility and comprehensive security have not been independently audited. Authentication and access depend on the configured services.
No independent runtime, installation, or comprehensive security audit was performed.
No installation, runtime, client matrix, or complete tool-contract audit was performed in this review.
Cloud pricing
Hosted endpoint details
Hosted deployment uses Arcade Cloud.
Standalone auth/resource-server configuration is operator-owned.

Original activity observations

2026-10-02T17:47:11Z: fix(arcade-core): describe every TypedDict field in tool schemas ( 950)

Technical identity

github.com/arcadeai/arcade-mcp

Exact implementation source: https://github.com/arcadeai/arcade-mcp ↗
Reviewed resource type: Framework
Repository identity: ArcadeAI/arcade-mcp
Package identity: pypi:arcade-mcp
Service endpoint: http://127.0.0.1:8000/docs

Evidence & Sources

Evidence basis: Recorded source evidence
Setup runtime verification: Not independently runtime-tested by MCP100 for this setup workflow.

Reviewed Oct 5, 2026 · Latest source observation Oct 6, 2026

GitHub repository ↗
arcadeai/arcade-mcp
Recorded repository material and project history.
Includes references whose attribution is not established.

Setup · Watchlist

Framework setup

Publisher instructions.

Source evidence observed Oct 13, 2026.

Additional setup details required — see the specific gaps below.

Install / connect
Publisher instructions
`uv tool install arcade-mcp` installs the Arcade framework CLI; `arcade new my_server` scaffolds YOUR OWN MCP server. Do not misrepresent the framework itself as an already hosted MCP endpoint. Start the generated server using its transport, register its actual endpoint/command, and test its sample tool.
uv tool install arcade-mcp
arcade configure claude
uv run server.py
Configure / authenticate
Publisher instructions
From the scaffold workspace use arcade configure claude or the documented cursor/vscode variant to register your custom server. OAuth provider workflows require their own app/consent and must not be implied by framework installation alone.
Check connection
Suggested check
Launch the publisher-created example server in the chosen transport, then inspect your own custom tool such as greet in the MCP client. Do not claim that an Arcade account has been authenticated based solely on this framework check.
Remove / uninstall / revoke
MCP100 setup guide
Stop the application-scaffolded MCP server process and remove its client registration. `arcade-mcp` builds MCP Apps; installing the CLI alone does not create a globally registered MCP server.

Official Documentation & Sources

What it does

Python framework for custom MCP tools, OAuth-aware authorization, evaluation and Arcade Cloud deployment.

Capability: Developer Tools
Resource type: Framework
Industry: Software & Technology
Best fit: Building custom authorized MCP servers and tool-evaluation workflows

Representative uses

Create custom internal tools
Declare OAuth scopes
Run stdio or remote server
Evaluate tool-call accuracy

Limitations

Hosted deployment uses Arcade Cloud.
Standalone auth/resource-server configuration is operator-owned.

Reviewed access

Access & risk

Review the exact access boundary before connecting this MCP to accounts, credentials, or sensitive data.

Authentication: OAuth/API keys/resource-server auth.
Permissions: Tool-declared upstream scopes.
Data handling: Hosted secrets are injected at runtime; custom tools decide data flow.
Cost: unknown
Risk and cautions: Framework risk is dominated by custom tools, requested scopes and deployment mode.

Access scope and cautions

Write actions: Depends on authored tools
Sensitive access: Upstream service data/credentials
Risk level: moderate

FAQ

Based on reviewed evidence as of 2026-10-06. Watchlist tracking is not a recommendation or a security certification.

Which tasks are documented for this implementation?

The reviewed tasks include: Create custom internal tools; Declare OAuth scopes; Run stdio or remote server; Evaluate tool-call accuracy.

What setup and authentication are documented?

Recorded installation/entrypoint step: uv tool install arcade-mcp. Recorded transport: stdio, sse, streamable_http. Prerequisites: Python/uv. These are the documented setup details, not a claim that installation or client compatibility was independently tested. Authentication: Tool-declared OAuth/API keys; Resource Server Auth for production HTTP.

Can it modify data or perform write operations?

Documented write actions include: Depends on authored tools. Recorded permission boundary: Tool-declared upstream scopes.

What data-handling boundary should I consider?

Hosted secrets are injected at runtime; custom tools decide data flow. Recorded sensitive-access context: Upstream service data/credentials.

What limitations and risks should I check before use?

Framework risk is dominated by custom tools, requested scopes and deployment mode. Hosted deployment uses Arcade Cloud. Standalone auth/resource-server configuration is operator-owned. No independent runtime, installation, or comprehensive security audit was performed.

Is this an official MCP?

Arcade AI is listed as First-party MCP in MCP100’s current evidence.