IMDb MCP server: let your AI agent search movies and TV shows by title

#What your agent gets

The tool takes query, a title or part of one, and an optional type. The default is any, which searches every kind of title. You can also pick one type: movie, tvSeries, tvMiniSeries, tvMovie, tvSpecial, short, video, videoGame, musicVideo or podcastSeries. Each result carries title, year, years, type, type_label, stars, popularity_rank, image and url. A search returns up to 8 matches.

In our check on 7 November 2026, a search for "the matrix" passed all 5 checks in one request and about 250 ms.

#How a request flows

1. You ask your agent: "Find movies called The Matrix on IMDb and show the year and the two lead actors."
2. The agent calls find_recipe, then run with imdb.com/search-titles and query: "the matrix" and type: "movie".
3. wield-mcp passes the run through the local bridge to the Wield extension.
4. The extension sends the one request from your browser, to the same host IMDb's search box calls.
5. The matches go back to your agent. Nothing passes through wield.run.

#What leaves your computer

The only host is v3.sg.media-imdb.com, the host IMDb's search box calls. A search makes one request. No login is needed. Passwords and cookies never leave your computer: Wield never reads, copies or sends them. Wield logs each request on your computer only, with method, host, path, status and timing, and never logs bodies or query values. Telemetry is off by default.

#IMDb's terms, in plain words

IMDb's Conditions of Use prohibit data mining, robots and screen scraping without written consent. The robots.txt on www.imdb.com disallows all bots. Wield runs this recipe only when you ask, one title lookup per call through IMDb's own suggestion service (v3.sg.media-imdb.com), from your own browser. It does not fetch title pages, ratings or bulk lookups. The scope is set: you start each search, one search per call, from your own browser, with no bulk collection and no purchases. This is not legal advice. Read IMDb's conditions yourself before you automate anything else.

#Limits

• Read only. The recipe searches. It cannot change anything on IMDb.
• One request per search. Ask for one title at a time, and do not loop it.
• No title pages. The result has no ratings, genres or plot. Its fields are the ones listed above.
• One type at a time. type takes a single value.
• Run limits. A run has at most 12 steps, 40 requests, 5 MB per response and 60 seconds.
• Bot checks stop the run. If a site shows a check, such as a CAPTCHA, the run stops and asks you to solve it in your own tab.

#Example prompts

• "Find IMDb titles matching 'breaking bad' and show the type and the two lead actors."
• "Search IMDb for movies called 'dune' from 2021 and give me the IMDb link."
• "List the IMDb matches for 'the matrix' and tell me which one has the highest popularity rank."

#FAQ

#Does the IMDb tool need a login?

No. The search needs no account, and the recipe does not ask for one.

#Which AI apps work with it?

Any MCP client. npx -y wield-mcp init sets up Claude Desktop, Claude Code, Cursor and Codex. Claude Desktop also has a one-click bundle on install.

#Can the tool rate a title or add it to a list?

No. It only reads search results. It cannot rate, save or change anything on IMDb.

#Why are ratings missing from the result?

Ratings and genres are not among the output fields. The result has the fields listed above, and nothing else.

#How many matches come back?

Up to 8, best match first. Your agent can sort them by popularity_rank if you ask.