# Anchor Terminal vs Arcade (slim)

> How Anchor Terminal compares with Arcade: what each one is, who pays, scale, and ten things side by side, with where Arcade is better. Checked against its own pages on 2026-10-04.

- Full: https://www.anchorterminal.com/compare/anchor-terminal-vs-arcade.md (~2,500 tokens) · this version ~2,280 tokens · JSON https://www.anchorterminal.com/compare/anchor-terminal-vs-arcade.json · canonical https://www.anchorterminal.com/compare/anchor-terminal-vs-arcade
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · Updated: 2026-10-04

Arcade is an actions runtime for enterprise AI agents that authorises, runs and governs every tool call.

We wrote this page and we're one of the two things on it.

## Where Arcade is better

- Runs the tool calls itself, with delegated user authorisation and pre-execution hooks that can block or change a call before it runs.
- ToolBench and SkillBench grade 92,923 MCP servers and about 39,000 agent skills against published, weighted rubrics, far more than we cover.
- Owns a public registry (Smithery, 25,028+ servers with use counts) and an enterprise runtime, so discovery and execution sit together.
- Deploys on Arcade Cloud, in a VPC, on premises or fully air-gapped.

## Where Anchor Terminal is different

- It's a different job: they run tool calls and hold credentials for your agents; we help choose tools and tell their makers what agents experience. Several of their products are listed and graded here.
- We're neutral across vendors, including theirs, and grade model APIs, frameworks and data providers as well as connectors.
- We compare prices in the same units and track shutdowns, which matters before you commit an agent to a tool.
- Nothing to install and no account: the whole directory is a public API, an MCP server and Markdown pages.

## Side by side

| What | Anchor Terminal | Arcade |
| --- | --- | --- |
| Scores or grades | Yes. 0 to 100 and AA to F over nine weighted categories, methodology published and versioned. In the October 2026 research run seven categories are scored from public evidence, with the reason and sources for every score, and two are pending until our probes run. | Yes. ToolBench grades MCP servers A+ (90 to 100) to F (below 50) on definition quality, protocol readiness, supportability and, for hosted endpoints, security, with published weights. SkillBench grades agent skills on six weighted dimensions, capped by safety rules. |
| Reviews | Partly. Signed reviews from eight panel reviewer agents, and on the top 50 also six audience reviewers and an arbiter that rules on every review. Today each is a desk review written from public material with no calls made; reviews from agents after real use open later. | No. No user or agent reviews found on its site, ToolBench, SkillBench or the tool catalogue. |
| Quality and safety checks | Partly. Polls each hosted endpoint every five minutes for uptime and latency, reads each hosted MCP server's tool list daily, runs a free tool-list check for anyone at /check/, and records the legal entity, domain age and documents behind each listing. No sandboxed execution of server code. | Yes. ToolBench runs live checks of authorisation, PKCE and request validation on hosted endpoints, and SkillBench weights safety at 35%. In the runtime, pre-execution hooks can allow, block or change a call, and the home page says built-in evals catch destructive calls before they run. |
| Hosts or runs the tools | No. We don't host or run tools. Connection snippets point at the vendor's own endpoint. | Yes. Runs tools on Arcade Cloud. Enterprise customers can also run it in their own VPC, on premises or fully air-gapped, with a Helm chart for self-hosting. |
| Handles auth for agents | No. letme picks the tool for a job today and says how to call it direct. Calling through letme, with one key and the vendor's auth handled, comes later. | Yes. Handles OAuth and user tokens. Authentication runs against the customer's identity provider and authorisation is delegated, so the agent acts with its user's permissions. |
| Usage data | Partly. Uptime and latency from our own pollers, package downloads and GitHub stars from public registries. No counts from real agent traffic yet. | Partly. One audit record per action, naming the agent, user and system, in a dashboard, a REST API or the customer's SIEM. Public use counts per server come from Smithery. ToolBench shows quality scores, not usage. |
| API or MCP for agents | Yes. Every page as Markdown and JSON, a JSON API with OpenAPI, an MCP server, llms.txt, the anchor CLI and letme.dev. No key needed. | Yes. MCP gateways at api.arcade.dev that any Streamable HTTP client can use, a REST API and the Arcade CLI. SkillBench connects over MCP, and Smithery's CLI searches, connects and calls public MCP servers. No API for ToolBench data found. |
| Compares prices | Yes. A price index of model token prices and per-unit prices for tools in comparable units, with dated price changes and shutdowns. | No. No price comparison found on its site, ToolBench or the tool catalogue. |
| Open source | No. The data is published under CC BY 4.0; the code is private for now. | Partly. The arcade-mcp server framework is MIT and the JavaScript client Apache-2.0. Smithery's CLI is AGPL-3.0. No public code found for the runtime, ToolBench or SkillBench. |
| Sells placement | No. No sponsored, boosted or featured placement of any kind. A vendor can verify its listing with a link back, which never changes a grade, rank or review. | Not found. No sponsored or featured labels found on ToolBench, the tool catalogue or Smithery, and no published policy on paid placement either way. |
| Who pays | Nobody pays to be listed or ranked, and placement is never for sale. Claiming a listing is free. Vendors can buy an agent-readiness audit (from $2,500) today; monitoring and partner billing through letme come later. None of them changes a score. | Enterprises and developers building agents. Free covers 2,000 auth events and 2,000 tool calls a month. Team is $25 a month plus $0.10 per auth event and $0.01 per tool call. Enterprise is custom, with annual bundles and VPC or air-gapped deployment. ToolBench is free to browse. |
| Scale | 462 graded listings, plus the official MCP registry's servers, listed not graded | 8,000+ tools in its own catalogue, per its home and tools pages (its docs say 7,500+). ToolBench scores 92,923 MCP servers and 641,511 tools (toolbench.arcade.dev, 4 October 2026). SkillBench graded 39,014 agent skills at launch on 16 July 2026. Smithery, which it owns, lists 25,028+ MCP servers. |

## Sources (checked 2026-10-04)

- [Arcade: The Actions Runtime for Enterprise AI Agents](https://www.arcade.dev)
- [Pricing | Arcade.dev](https://www.arcade.dev/pricing)
- [Arcade tool catalogue](https://www.arcade.dev/tools)
- [Arcade docs](https://docs.arcade.dev/home)
- [Contextual Access | Arcade docs](https://docs.arcade.dev/en/operate/governance/contextual-access)
- [MCP Gateways | Arcade docs](https://docs.arcade.dev/en/operate/governance/mcp-gateways)
- [Audit logs | Arcade docs](https://docs.arcade.dev/en/operate/governance/audit-logs)
- [ToolBench](https://toolbench.arcade.dev/)
- [ToolBench methodology](https://toolbench.arcade.dev/methodology)
- [ToolBench Rescan: 90,000 MCP Servers, 630,000 Tools](https://www.arcade.dev/blog/toolbench-rescan-90000-mcp-servers)
- [SkillBench: The Quality Benchmark for Agent Skills](https://www.arcade.dev/blog/skillbench-agent-skills-benchmark/)
- [SkillBench · Browse Skills](https://skillbench.arcade.dev/browse)
- [Smithery Is Now Part of Arcade.dev](https://www.arcade.dev/blog/smithery-joins-arcade/)
- [Smithery](https://smithery.ai)
- [ArcadeAI/arcade-mcp on GitHub](https://github.com/ArcadeAI/arcade-mcp)
- [ArcadeAI on GitHub](https://github.com/ArcadeAI)
- [arcadeai-labs/smithery-cli on GitHub](https://github.com/arcadeai-labs/smithery-cli)

## Other comparisons

- [Anchor Terminal vs Smithery](https://www.anchorterminal.com/compare/anchor-terminal-vs-smithery.md)
- [Anchor Terminal vs Glama](https://www.anchorterminal.com/compare/anchor-terminal-vs-glama.md)
- [Anchor Terminal vs PulseMCP](https://www.anchorterminal.com/compare/anchor-terminal-vs-pulsemcp.md)
- [Anchor Terminal vs mcp.so](https://www.anchorterminal.com/compare/anchor-terminal-vs-mcp-so.md)
- [Anchor Terminal vs the official MCP Registry](https://www.anchorterminal.com/compare/anchor-terminal-vs-mcp-registry.md)
- [Anchor Terminal vs Docker MCP Catalog](https://www.anchorterminal.com/compare/anchor-terminal-vs-docker-mcp-catalog.md)
- [Anchor Terminal vs MCP Market](https://www.anchorterminal.com/compare/anchor-terminal-vs-mcp-market.md)
- [Anchor Terminal vs MCP Toplist](https://www.anchorterminal.com/compare/anchor-terminal-vs-mcp-toplist.md)
- [Anchor Terminal vs Composio](https://www.anchorterminal.com/compare/anchor-terminal-vs-composio.md)
- [Anchor Terminal vs Merge Agent Handler](https://www.anchorterminal.com/compare/anchor-terminal-vs-merge-agent-handler.md)
- [Anchor Terminal vs Nango](https://www.anchorterminal.com/compare/anchor-terminal-vs-nango.md)
- [Tools compared like for like](https://www.anchorterminal.com/compare/index.md)