PixVerse MCP Server
November 20, 2026
Table of contents
- Endpoint and authorization
- Setup
- Use it in the Claude app or ChatGPT
- Tools
- Long-running work
- Example prompts
- Limits and notes
The useapi.net PixVerse MCP server lets an AI assistant such as the Claude app, ChatGPT, Claude Code, Cursor or VS Code Copilot make videos, images, music and speech with PixVerse on your own PixVerse accounts. It is a hosted Model Context Protocol server with 38 tools: generate videos with PixVerse’s own models and third-party ones such as Seedance 2.5 from text, an image, an effect template, first and last frames, keyframes or reference images, then extend, modify, lip-sync or upscale them, animate a character image with motion control, generate and upscale images, make songs, music videos and speech, upload your own files and clean up your PixVerse history. Each tool calls the PixVerse API, so the same rules and limits apply. Nothing needs to be installed or hosted.
Is there a PixVerse MCP server? Yes. useapi.net, a third-party service not affiliated with PixVerse, hosts one at https://api.useapi.net/v2/pixverse/mcp with 38 tools for PixVerse video, effects, images, music and speech on your own PixVerse subscription, nothing to install.
Prefer REST? Code: github.com/useapi/pixverse-api
See all useapi.net MCP servers on Model Matrix & MCP.
Endpoint and authorization
https://api.useapi.net/v2/pixverse/mcp
- Transport: Streamable HTTP.
- The path is
/v2/, because the server sits on version 2 of the PixVerse API. The other useapi.net MCP servers use/v1/. - Authorization: your useapi.net API token in the
Authorization: Bearer …header, see Setup useapi.net. The Claude app and ChatGPT sign in with OAuth instead, see Use it in the Claude app or ChatGPT. - You need at least one PixVerse account added with Setup PixVerse. Adding and removing accounts is not available over MCP: use POST accounts/email or the setup page.
Setup
Replace YOUR_API_TOKEN with your useapi.net API token.
Claude Code
claude mcp add --transport http useapi-pixverse https://api.useapi.net/v2/pixverse/mcp --header "Authorization: Bearer YOUR_API_TOKEN"
Cursor
In ~/.cursor/mcp.json:
{
"mcpServers": {
"useapi-pixverse": {
"url": "https://api.useapi.net/v2/pixverse/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_TOKEN"
}
}
}
}
VS Code
In .vscode/mcp.json:
{
"servers": {
"useapi-pixverse": {
"type": "http",
"url": "https://api.useapi.net/v2/pixverse/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_TOKEN"
}
}
}
}
Windsurf and other clients
Any client that supports remote MCP servers over HTTP with custom headers works. The usual JSON shape:
{
"mcpServers": {
"useapi-pixverse": {
"serverUrl": "https://api.useapi.net/v2/pixverse/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_TOKEN"
}
}
}
}
Some clients name the address field url instead of serverUrl.
Test it with curl
curl -s https://api.useapi.net/v2/pixverse/mcp \
-H "Authorization: Bearer YOUR_API_TOKEN" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl","version":"1"}}}'
curl -s https://api.useapi.net/v2/pixverse/mcp \
-H "Authorization: Bearer YOUR_API_TOKEN" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
The first call answers with the server’s name useapi-pixverse and its instructions for the assistant, the second with the 38 tools.
Use it in the Claude app or ChatGPT
The Claude app (claude.ai and Claude Desktop) and ChatGPT can add this server as a custom connector. They sign in with OAuth: you give the app the URL, it opens a useapi.net page, and you paste your useapi.net API token there once. Both work on their Free plans. The API-token setups above for Claude Code, Cursor, VS Code and other clients keep working as before.
https://api.useapi.net/v2/pixverse/mcp
Claude app
- Open Settings → Connectors and click Add custom connector.
- Enter a name, for example
useapi PixVerse, and the URL above, then click Add. - Click Connect. Claude opens the useapi.net sign-in page.
- Paste your useapi.net API token and click Allow.
Claude asks before each tool call unless you choose Always allow. The Free plan allows one custom connector.
ChatGPT
- Open Plugins, click Add and choose Add custom MCP server.
- Enter a name and the URL above, and set Authentication to OAuth.
- Tick ChatGPT’s “I understand and want to continue” box, then click Create as a plugin.
- Click Connect. ChatGPT opens the useapi.net sign-in page.
- Paste your useapi.net API token and click Allow.
ChatGPT asks you to confirm each tool call that creates or changes something. The Free plan works without developer mode (as of October 25, 2026).
How the sign-in works
- Your API token is in your welcome email, see Setup useapi.net. The useapi.net page links there too.
- Each connection is for one MCP server. Connect each server you want to use on its own.
- The app gets an access token that lasts 1 hour and renews it on its own. A connection you don’t use for 90 days lapses, and you connect again.
- Resetting your API token disconnects every app connected with it.
- Results come back as public links, so they open right in the app.
Tools
Accounts are named by their email. Tools that start a job take an optional email, and without it a random account with a free slot is picked. A job on an existing video or image (video_id, image_id) runs on that video’s or image’s account, because the id names it. Files uploaded with pixverse_upload_file work on any of your accounts. Tools that read one account’s lists need email only when more than one account is added. Every create tool also takes maxJobs (1 to 8) to override the account’s jobs at once for that request, and wait.
Accounts and files
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_list_accounts | Lists your accounts as stored: email, maxJobs, session expiry and any pending account error, without calling PixVerse. | |
pixverse_get_account | One account’s stored settings: email, maxJobs, the session’s expiry and any pending error, without calling PixVerse. | email |
pixverse_get_features | Reads an account’s PixVerse plan live: plan name, credit balances, refresh and expiry dates, generations at once, and the image models and video qualities the plan allows. | email |
pixverse_upload_file | Uploads your own image (PNG, JPEG, GIF or WebP), video (MP4, MOV, WebM and more) or audio (WAV or MP3), up to 50 MB, from a public https URL or sent as contentBase64. Returns the file’s path for the *_path parameters of the other tools. Uploads cannot be deleted. | url or contentBase64, mimeType, email |
The parameters match GET accounts, GET accounts/email, GET features and POST files.
Images
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_create_image | Generates 1 to 4 images with Qwen Image (the default), Nano Banana, Seedream, GPT Image or Kling image models, from a prompt and up to 16 reference images, by model. In Relax Mode it is free where the plan includes the model. The first image comes back as an image the assistant can see. | prompt, model, quality, aspect_ratio, detail_level, create_count, image_paths, seed, relax_mode, email |
pixverse_get_image | Reads one image, and with wait: true waits for it to finish. A finished image comes with its image_url, size and credits used, and as an image block. | image_id, wait |
pixverse_list_images | Lists one account’s images, including those still queued or generating and those made on pixverse.ai, page by page. | email, offset, limit |
pixverse_upscale_image | Upscales a PixVerse image or your own upload with PixVerse Upscale or Topaz Gigapixel, up to 4K. quote: true answers the exact price in credits and creates nothing. | image_id or image_path, model, quality, quote |
pixverse_delete_image | Deletes one image from its PixVerse account. It cannot be undone. | image_id |
The parameters match POST images/create, GET images/image_id, GET images, POST images/upscale and DEL images/image_id. The numbered reference images of the REST API (image_path_1, image_path_2 and so on) are the list image_paths here.
Videos
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_create_video | Generates a video from text, an image or an effect template with PixVerse V6 (the default), C1, V5.x or third-party models such as Seedance 2.5, Kling 4.0, Veo 3.1 and MiniMax H3. Third-party models need a Pro plan or higher. | prompt, model, duration, quality, aspect_ratio, first_frame_path, template_id, template_images, audio, multi_shot, preview_mode, off_peak_mode, relax_mode, email |
pixverse_create_video_frames | Generates a video that moves from a first frame to a last frame, with V5.6 (the default), V5.5, V5, V6, C1 or third-party models. | first_frame_path, last_frame_path, prompt, model, duration, quality, audio, off_peak_mode |
pixverse_create_transition | Generates one video through 2 to 7 keyframes in order, each move with its own prompt and duration, with V5, V5.5 or V5.6 (3 or more frames on V5 only). | frames, model, quality, audio, off_peak_mode |
pixverse_create_video_fusion | Generates a video from reference images, and on the models that take them reference videos and audio, cited in the prompt as @image1, @video1 or @audio1 (PixVerse’s Reference mode). | prompt, image_paths, video_paths, audio_paths, model, duration, quality, aspect_ratio, audio |
pixverse_create_final_video | Renders a finished seedance-2.5-draft video again with Seedance 2.5 at 1080p or 4K, while the draft is valid (drafts are kept 14 days). | video_id, quality, relax_mode |
pixverse_extend_video | Continues a PixVerse video or your own upload from its last frame with V6 or Grok Imagine. | video_id or video_path, model, prompt, duration, quality, audio |
pixverse_modify_video | Changes what is in a PixVerse video or your own upload by a prompt, with V5.5 at 360p, 540p or 720p. | prompt, video_id or video_path, quality |
pixverse_motion_control | Makes the person or animal in a character image move like the subject of a reference video, with V5.6 and no prompt. | image_path, video_path, quality |
pixverse_lipsync_video | Makes the person in a video speak, from your own audio of up to 60 seconds or from text read by a lip-sync voice, with V5. | video_id or video_path, audio_path or prompt with speaker_id, original_sound_switch |
pixverse_upscale_video | Upscales a PixVerse video or your own MP4 or MOV upload of up to 3 minutes with PixVerse Upscale or Topaz Proteus, Starlight Fast or Starlight Precise, to 1080p or 4K. quote: true answers the exact price in credits and creates nothing. | video_id or video_path, model, quality, quote |
The parameters match POST videos/create, POST videos/create-frames, POST videos/create-transition, POST videos/create-fusion, POST videos/create-final, POST videos/extend, POST videos/modify, POST videos/motion-control, POST videos/lipsync and POST videos/upscale. Durations, qualities and ratios of each model are on Model capabilities.
The numbered fields of the REST API are lists here:
image_paths,video_pathsandaudio_pathsofpixverse_create_video_fusionstand forframe_N_path,video_N_pathandaudio_N_path. The first item ofimage_pathsis@image1in the prompt, the second@image2.template_imagesofpixverse_create_videoholds up to 5 images for a multi-input effect template, in place offrame_1_pathtoframe_5_path.framesofpixverse_create_transitionis a list of keyframes, each with itspathand thepromptanddurationof the move to the next frame, in place offrame_N_path,prompt_N_to_Mandduration_N_to_M.
Video reads, effects and voices
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_get_video | Reads one video from any create, extend, modify, lip-sync or upscale tool, and with wait: true waits for it to finish. A finished video comes with its url, size, duration and model, and a Seedance 2.5 draft says whether it can still be rendered in full and when it expires. | video_id, wait |
pixverse_list_videos | Lists one account’s videos, including those still queued or generating, music videos and those made on pixverse.ai, page by page. | email, offset, limit |
pixverse_delete_video | Deletes one video or music video from its PixVerse account. It cannot be undone. | video_id |
pixverse_list_effects | Lists PixVerse’s more than 900 effect templates for pixverse_create_video, with each template’s id, name, number of input images, model, qualities, duration and example prompt. search finds them by name or prompt. | search, offset, limit |
pixverse_list_lipsync_voices | Lists the voices for lip-sync speech, with a sample of each, for pixverse_lipsync_video and the V5 lines of pixverse_create_video. | email |
The parameters match GET videos/video_id, GET videos, DEL videos/video_id, GET videos/effects and GET videos/voices.
Music and music videos
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_create_music | Generates a song or an instrumental from a prompt with MiniMax Music 2.6 (the default), ElevenLabs Music v1 or Google Lyria 3 Pro: your own lyrics on Music 2.6 and v1, or lyrics the model writes, and up to 10 reference images on Lyria 3 Pro. | prompt, model, lyrics, instrumental, image_paths, email |
pixverse_get_music | Reads one track, and with wait: true waits for it to finish. A finished track comes with its mp3 url, duration and model. | audio_id, wait |
pixverse_list_music | Lists one account’s music tracks, page by page. | email, offset, limit |
pixverse_create_music_video | Makes a full music video from a song (an uploaded MP3 of 10 seconds to 6 minutes, or a library song) and optionally a character photo: performance, expressive or storytelling mode, a style, lyric captions, 720p or 1080p. The first music video of every account is free. quote: true answers the exact price and creates nothing. | audio_path, mode, style, image_path, subtitles, lyrics, prompt, aspect_ratio, quality, quote |
pixverse_get_music_video | Reads one music video, and with wait: true waits for it to finish. A finished one comes with the video and extra files such as the copy without captions. | video_id, wait |
pixverse_list_music_videos | Lists one account’s music videos, page by page. | email, offset, limit |
pixverse_list_music_video_library | Lists PixVerse’s built-in songs, characters and styles for pixverse_create_music_video, for when you have no song or photo of your own. | type |
The parameters match POST music/create, GET music/music_id, GET music, POST music-video/create, GET music-video/video_id, GET music-video and GET music-video/library. The numbered reference images of POST music/create (image_path_1 to image_path_10) are the list image_paths here.
Speech
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_create_speech | Turns text into speech with MiniMax Speech 2.8 HD (the default) or Turbo, or ElevenLabs Multilingual v2, v3 or Turbo v2.5, with voice settings such as speed, pitch and emotion on MiniMax, and stability and similarity on ElevenLabs. | text, voice_id, model, language_code, speed, volume, pitch, emotion, stability, similarity_boost |
pixverse_get_speech | Reads one speech clip, and with wait: true waits for it to finish. A finished clip comes with its audio url, duration, voice and language. | audio_id, wait |
pixverse_list_speech | Lists one account’s speech clips, page by page. | email, offset, limit |
pixverse_list_speech_voices | Lists the text-to-speech voices of one model’s provider, with gender, age, accent, languages and a preview, page by page. language narrows them to voices for that language. | model, language, offset, limit |
pixverse_list_speech_models | Lists the text-to-speech models live from PixVerse, with each model’s character limit, languages and credit formula. | email |
The parameters match POST speech/create, GET speech/speech_id, GET speech, GET speech/voices and GET speech/models.
Running jobs
| Tool | What it does | Main parameters |
|---|---|---|
pixverse_list_running_jobs | Lists the videos, images, music and speech the API is tracking now and, per account, maxJobs with its running and free slots. Jobs started on pixverse.ai or in Off-Peak mode are not tracked. | |
pixverse_cancel_job | Stops tracking a running job and frees its slot on the account. It does not stop the job at PixVerse. | id |
The parameters match GET scheduler/available and DEL scheduler/id.
The full input schema of every tool comes from tools/list.
Long-running work
MCP clients usually give up on a tool call after about a minute, so a tool waits up to about 50 seconds:
- A quick result can come back from the create tool itself.
- Otherwise the tool returns the id, and the matching get tool (
pixverse_get_video,pixverse_get_image,pixverse_get_music,pixverse_get_speechorpixverse_get_music_video) withwait: truepicks it up. Ask the assistant to keep checking until it is done. - A job in Off-Peak mode (
off_peak_mode: true) is not waited for: the tool returns itsvideo_idat once, andpixverse_get_videoreads it later.
Clients that ask for progress notifications get one about every 10 seconds while a tool waits.
Videos, images, music and speech come back as public media.pixverse.ai links.
Example prompts
- “Make a 5-second PixVerse V6 video at 1080p with sound of a paper boat drifting down a rainy street at night.”
- “Find a PixVerse effect with ‘hug’ in its name and use it on this photo.”
- “Upload these two photos and make a video that moves from the first one to the second.”
- “Make a Seedance 2.5 draft of this prompt, and if it looks right, render it in 4K.”
- “Generate an image of a cozy mountain cabin in winter, then upscale it to 4K.”
- “How many credits would it cost to upscale this video to 4K with Topaz Starlight Fast?”
- “Make a music video from this MP3 with my photo as the performer, 9:16, with lyric captions.”
- “Read this paragraph aloud with an ElevenLabs v3 voice.”
- “How many credits are left on my PixVerse account?”
Limits and notes
- The tools use your PixVerse accounts and the same rules as the REST API, such as
maxJobsand the models of each plan. Generations use the account’s PixVerse credits, see Pricing, andpixverse_get_featuresshows what is left. Upscales and music videos takequote: truefor a free exact price. - A 412 means the plan or the credits do not allow that job. A 429 means your accounts are at
maxJobs: checkpixverse_list_running_jobsand retry later. - A client may send up to 20 JSON-RPC requests in one batch.
- If the client sends an
MCP-Protocol-Versionheader, it must be one of the supported versions:2025-11-25,2025-06-18,2025-03-26or2024-11-05. - Tool calls show in your usage stats (
bot=pixverse) through the REST call each tool makes, the same rows a direct API request writes.