Internal reference report
Firecrawl
Anonymous endpoint profile
4
Metadata checks passed
3
Advertised tools
- Tool executions
- 0
- ChatGPT / Claude
- Not tested
- OAuth
- Not tested
Internal, unpaid reference. GraphCanon’s owner works at Firecrawl.
MCP reviews · Evidence first
MCP reviews you can inspect, not just trust.
Compare recorded evidence for a specific endpoint and profile. Start with the available metadata baselines, then see the deeper testing we plan to add.
A report before a rating.
Four passing metadata checks do not make an 80/100 score. We show the checks, the evidence and the gaps.
Inspect the comparisonCurrent test coverage
MCPMetadata available
ChatGPTPlanned · not tested
ClaudePlanned · not tested
01 / Available now
Two endpoints. The same bounded checks.
These are internal reference runs, not independent reviews or customer endorsements.
Internal reference report
GraphCanon
Anonymous endpoint profile
4
Metadata checks passed
9
Advertised tools
- Tool executions
- 0
- ChatGPT / Claude
- Not tested
- OAuth
- Not tested
Internal, unpaid reference. GraphCanon is our own product.
Swipe or scroll sideways to compare both endpoints.
| Current test coverage | Firecrawl | GraphCanon |
|---|---|---|
| Metadata checks passed | 4 passed | 4 passed |
| Advertised tools | 3 | 9 |
| Tool executions | 0 | 0 |
| OAuth / ChatGPT / Claude | Not tested | Not tested |
| Overall score | Not assigned | Not assigned |
Recorded 15 October 2026 using runner 1.0.1. Different tool counts reflect different profiles, not a quality ranking. These snapshots do not monitor current availability.
02 / How we review
Each claim needs its own kind of evidence.
01 / Available now
Inspect the connection
Protocol negotiation, initialization, the advertised tool catalog and SDK metadata shape. The current reports stop here.
02 / Planned
Exercise real tasks
Scoped sandbox tasks, input errors, authentication and recovery. Tool declarations alone cannot establish behavior or safety.
03 / Planned
Test in the actual client
Separate ChatGPT and Claude trials with the product, model, permissions and date recorded. API tests and emulation will not count as real-client verification.
03 / Review options
Pay for useful testing, never for a better result.
The public baselines are free. Paid review scope and pricing are still being developed; no paid service is available to purchase here.
Free · available
Public baselines
Read the two published metadata reports and download their evidence. No account is required.
Inspect the comparisonPlanned
Scoped MCP review
A planned specialist review of one endpoint, agreed tasks and client configurations, with findings you can reproduce.
Scope and price to be validated.
Planned
Release monitoring
Planned repeat checks against an agreed baseline, with changes and newly missing coverage made visible.
Scope and price to be validated.
Have an MCP endpoint worth reviewing?
Email a public endpoint or repository, the client you care about and one task you need to work. We will discuss scope and sandbox access before any testing.
Opens your email app. Do not send keys, tokens or private customer data. This is a scope request, not an automatic scan or purchase.
Payment will not change findings or organic rankings. Internal relationships are disclosed on every reference report. No official certification or platform approval is implied.
Explore the AI tool catalog