MCP
View as Markdown · Open in Claude Desktop
Neattrim speaks the Model Context Protocol, so an assistant can run the same tools the API offers. MCP is open to every account within its allowance, like the API. The tool descriptions are written for the model and come from the same schema file as the API, so names and fields match.
If you are not a developer, see the assistants page.
Servers
| Server | Address | Sign-in |
|---|---|---|
| core | https://api.neattrim.com/mcp |
OAuth 2.1 |
| generate | https://api.neattrim.com/generate/mcp |
OAuth 2.1, scope generate.voice |
| full | https://api.neattrim.com/full/mcp |
OAuth 2.1 or API key |
| local | npx -y @neattrim/media |
neattrim login |
Cloud servers sign in with OAuth 2.1. The full server also takes an API key. The local server runs on your own computer and needs no sign-in for device tools; neattrim login or NEATTRIM_API_KEY unlocks cloud tools.
Connect your client
The hosted servers need no install: add the address to your client and sign in with your Neattrim account when it asks. Use https://api.neattrim.com/mcp (the core server) unless you need speech, dubbing and generated video (https://api.neattrim.com/generate/mcp), or one server with every tool (https://api.neattrim.com/full/mcp). In each example below, swap the address to use another server.
| Server | Address | Cursor | VS Code |
|---|---|---|---|
| Core | https://api.neattrim.com/mcp |
Add to Cursor | Add to VS Code |
| Generate | https://api.neattrim.com/generate/mcp |
Add to Cursor | Add to VS Code |
| Full | https://api.neattrim.com/full/mcp |
Add to Cursor | Add to VS Code |
Claude and Claude Desktop
In Claude, open Customize, then Connectors, choose Add custom connector, enter Neattrim and https://api.neattrim.com/mcp, and select Add. Claude opens a Neattrim window to sign in and allow the connection. On a Team or Enterprise plan, an owner of the organisation adds the connector in the organisation settings first. A free Claude account can add one custom connector.
Claude Code
claude mcp add --transport http neattrim https://api.neattrim.com/mcp
Then run /mcp in a session to sign in. Add -s user to use it in every project, or -s project to share it through the project's .mcp.json.
Cursor
Select Add to Cursor above, or add the server to ~/.cursor/mcp.json (every project) or .cursor/mcp.json (one project):
{ "mcpServers": { "neattrim": { "url": "https://api.neattrim.com/mcp" } } }
Cursor asks you to sign in the first time a tool runs.
VS Code
Select Add to VS Code above, or add the server to .mcp.json in your workspace (or ~/.copilot/mcp-config.json for every workspace):
{ "mcpServers": { "neattrim": { "type": "http", "url": "https://api.neattrim.com/mcp", "tools": ["*"] } } }
From a terminal: code --add-mcp '{"name":"neattrim","type":"http","url":"https://api.neattrim.com/mcp"}'. VS Code asks you to sign in when the server starts.
Windsurf (Devin Desktop)
Add the server to ~/.codeium/windsurf/mcp_config.json, then refresh the MCP list and sign in when asked:
{ "mcpServers": { "neattrim": { "url": "https://api.neattrim.com/mcp" } } }
ChatGPT
In ChatGPT on the web, open the plugins page, select the plus button and Add custom MCP server. Enter Neattrim and the server address https://api.neattrim.com/full/mcp, choose OAuth, confirm, create it as a plugin and install it. ChatGPT opens a Neattrim window to sign in. Which ChatGPT plans can add a custom server is set by OpenAI.
Codex CLI
codex mcp add neattrim --url https://api.neattrim.com/mcp
codex mcp login neattrim
This writes the server to ~/.codex/config.toml:
[mcp_servers.neattrim]
url = "https://api.neattrim.com/mcp"
Gemini CLI
gemini mcp add -s user --transport http neattrim https://api.neattrim.com/mcp
Then run /mcp auth neattrim in a session to sign in. In ~/.gemini/settings.json the same server is:
{ "mcpServers": { "neattrim": { "httpUrl": "https://api.neattrim.com/mcp" } } }
Run the local server
Not on npm yet. The local server
@neattrim/mediais built and tested but not published: it arrives with the first client release, listed in the changelog. The hosted servers above work today and need no install.
npx -y @neattrim/media
Or run neattrim mcp <folder> from the CLI to limit it to one folder.
Tools
| Tool | What it does | Sets |
|---|---|---|
get_capabilities |
Lists what this server can do: tools, accepted formats, per-file limits for the signed-in account, how much of the account's monthly allowance is left, and how to install the local version that runs on the user's own computer. | core, generate, full, local_free |
estimate_usage |
Estimates how much of the account's allowance a planned job will use, from the media duration and options, and whether it fits what is left. | core, generate, full, local_free |
create_upload |
Gets a place to put a file that is on the user's device or in your sandbox when there is no public URL. | core, generate, full |
get_job |
Gets the status and results of a job started by another tool. | core, generate, full, local_cloud |
cancel_job |
Stops a running job. | core, generate, full, local_cloud |
list_library |
Lists the user's recent results in their library, newest first, with names, sizes and deletion dates. | core, generate, full, local_cloud |
delete_file |
Permanently deletes an uploaded file or a result now, before its scheduled deletion. | core, generate, full, local_cloud |
probe_media |
Reads a media file's container, codecs, duration, resolution, frame rate, size and audio tracks without changing it. | core, full, local_free |
trim_video |
Cuts one segment out of a video or audio file. | core, full, local_free |
compress_video |
Makes a video file smaller. | core, full, local_free |
convert_media |
Changes a file's container or codec, extracts the audio track to MP3, M4A or WAV, or removes the audio from a video. | core, full, local_free |
edit_video |
Applies simple edits to a video in the order given: crop (by aspect ratio or pixels), rotate or flip, merge several clips, change speed, reverse, loop, add a music track the user supplies, or add text. | core, full, local_free |
make_gif |
Turns a video clip of up to 30 seconds, or a set of images, into an animated GIF, optionally with caption text at the top and bottom. | core, full, local_free |
make_meme |
Adds meme-style caption text to an image, GIF or short video in a classic top and bottom layout or a caption band. | core, full, local_free |
edit_audio |
Cuts, fades, changes volume or normalises loudness of an audio file, or of the sound of a video; can make an M4R ringtone. | core, full, local_free |
edit_image |
Converts (HEIC, JPG, PNG, WebP), compresses, or resizes images by pixels, percent or a target file size in KB; accepts up to 50 images at once. | core, full, local_free |
convert_subtitles |
Converts a subtitle file between SRT, VTT, ASS and plain text, keeping timings; can shift all timings. | core, full, local_free |
transcribe_media |
Converts speech in a video or audio file to text in 99 languages: plain text, timestamped text, SRT, VTT or word-level JSON; optionally labels speakers. | core, full, local_free, local_cloud |
translate_subtitles |
Translates a subtitle file or a transcript into up to 6 languages, keeping timings; can apply a glossary of terms that must stay fixed and run a quality-check pass that flags doubtful lines. | core, full, local_cloud |
review_translation |
Checks a translated subtitle file against its source: untranslated lines, glossary terms, numbers, tags, empty or duplicate lines, reading speed and line length, plus an optional language review pass. | core, full, local_free, local_cloud |
burn_captions |
Draws a subtitle file onto a video in a named caption style, producing a new video. | core, full, local_free |
subtitle_video |
Creates subtitles for a video or audio file from its speech, optionally translated. | core, full, local_cloud |
make_notes |
Turns a lecture, meeting or podcast recording into sectioned notes with timestamps; can also produce a quiz, flashcards (Anki CSV), chapters, titles and show notes. | core, full, local_cloud |
clean_audio |
Removes background noise and room echo from speech so it sounds like a studio recording. | core, full, local_free, local_cloud |
remove_music |
Removes background music from a video or audio file and keeps the voices. | local_free |
separate_vocals |
Splits a song or recording into two files: vocals and accompaniment (karaoke track). | local_free |
remove_image_background |
Removes the background from images, returning transparent PNGs. | core, full, local_free |
make_passport_photo |
Turns one portrait into a passport, visa, ID or exam photo: finds the face, crops to the chosen country or exam rule, sets the background and meets the pixel size and KB limit. | core, full, local_free |
clean_up_talking_video |
For talking-head videos: removes filler words, retakes and long silences, optionally improves the voice and labels speakers. | core, full, local_free, local_cloud |
process_podcast |
Prepares a podcast episode: cleans the sound, makes a transcript with speakers, show notes, chapters, titles and a description. | core, full, local_cloud |
smart_edit_video |
Edits a video from its speech: removes fillers, repeats and retakes, cuts it down to a target length, or follows a short instruction. | core, full, local_cloud |
make_shorts |
Finds 3 to 5 strong moments in a long video and turns each into a vertical 9:16 clip with captions. | core, full, local_cloud |
reframe_video |
Changes a video's aspect ratio (for example 16:9 to 9:16 or 1:1) and keeps the speaker or subject in frame. | core, full, local_free |
list_voices |
Lists voices for text to speech and dubbing by language, gender and tier. | generate, full, local_cloud |
text_to_speech |
Reads text aloud with a chosen voice, returning MP3 or WAV. | generate, full, local_cloud |
dub_video |
Translates the speech of a video into another language and replaces it with a synthetic voice, keeping the background sound. | generate, full, local_cloud |
text_to_video |
Turns a script, or an audio file, into a captioned video in 9:16, 1:1 or 16:9 with a synthetic voice and visuals. | generate, full, local_cloud |
generate_reel |
Writes a short hook-first script on a topic and renders it as a vertical captioned video, with a suggested caption and hashtags. | generate, full, local_cloud |
Voice cloning is web-only. Music and vocal separation and the studio clean-up preset run only on your own computer. get_capabilities lists what a server cannot do, with the reason.