FrameThrower is the world's largest library of film stills for inspiration, learning, and making movies faster. Find movie screenshots and cinematography references with AI-powered visual search. Explore lighting, composition, colour and shot size, then organise your references into lookbooks and export PDF boards for pitches and production. The FrameThrower MCP server connects Claude, ChatGPT, Cursor and other MCP clients to film references through four tools: search_frames, find_by_craft, find_similar and get_frame_details. Search by mood, composition and cinematographic craft, discover similar frames, and retrieve frame metadata and links. Connect using OAuth 2.1 at https://framethrower.ai/api/mcp . Setup and documentation: https://framethrower.ai/mcp . Included in the free plan, with usage metered against your credit balance.

MCP readiness

B · 83/100

Readiness is how fully this listing follows its protocol and how it behaved in our probes over 30 days. It is shown for reference and does not move it in search. Search order and leaderboards use the separate quality grade (here B, 74.8/100), which weighs availability, conformance, p95 speed and descriptor. How the two differ

Valid MCP handshake	10/10	initialize and tools/list both answered.
Live MCP endpoint	25/25	Endpoint answers the MCP initialize handshake (1689 ms).
Protocol revision	8/10	Speaks MCP 2025-11-25, 1 revision(s) behind 2026-07-28.
Signed descriptor	n/a	No MCP standard defines a signed server descriptor.
Uptime (30 days)	15/15	30/30 probes succeeded over 30 days (100%).
Tool declaration	10/10	Lists 17 tool(s) with descriptions.
Verified identity	5/10	Publisher declared (repository, package or homepage); claim this listing to verify domain control.
Freshness + modern flags	0/5	not seen in a source for over 30 days.
Auth declaration	2/5	Declares credential scheme: smithery_api_key.

Applicable criteria total 90 points, scaled to 100.

Responsiveness: Responsive (1689 ms; bands ≤2.5 s, ≤5 s, ≤8 s, slower; penalty only).

Measured 2026-11-11.

Nine criteria, the same as the A2A readiness score, read from our daily MCP probe (initialize + tools/list) and 30-day probe history. No MCP standard defines a signed server descriptor, so that criterion is n/a and the other 90 points are scaled to 100. Grades: A>=90, B>=75, C>=60, D>=40, F<40. Display only: search order uses the quality grade.

Compared with peers

Against 53,039 reachable MCP servers in our daily snapshot (2026-11-12 06:46 UTC). Display only; it does not change search order.

Readiness: Higher than 60% of peers; This listing 83 · peer median 81
Latency: Faster than 34% of peers; This listing 470 ms · peer median 326 ms

Where it loses the most points, and how to win them back

Verified identity −5 — Claim the listing and prove control of its domain.
Now: Publisher declared (repository, package or homepage); claim this listing to verify domain control.
Freshness + modern flags −5 — Keep the listing current in the registry or source it comes from.
Now: not seen in a source for over 30 days.
Auth declaration −3 — State how to authenticate (an auth scheme, or that none is needed).
Now: Declares credential scheme: smithery_api_key.

How many agent tools actually work · All statistics

Listed in

Directories this entry was found in.

smithery since 2026-09-07 ↗

How to connect

MCP endpoint (streamable-http)

https://server.smithery.ai/leo-5n21/framethrower/mcp

JSON-RPC initialize probe

curl -X POST https://server.smithery.ai/leo-5n21/framethrower/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}'

Homepage

https://framethrower.ai

Listed at (smithery)

https://framethrower.ai