Higgsfield
MCP server for Higgsfield AI cinematic image and video generation
Ask AI about Higgsfield
Powered by Claude Β· Grounded in docs
I know everything about Higgsfield. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
Higgsfield MCP
MCP server for Higgsfield AI β generate images and videos directly from Claude or any MCP-compatible client.
Setup
Node.js 18+ required. Get your API keys at https://cloud.higgsfield.ai/api-keys.
Claude Desktop
Edit %APPDATA%\Claude\claude_desktop_config.json (Windows) or ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"higgsfield": {
"command": "npx",
"args": ["-y", "higgsfield-mcp"],
"env": {
"HF_API_KEY": "your_api_key",
"HF_SECRET": "your_secret"
}
}
}
}
Claude Code
claude mcp add higgsfield -- npx -y higgsfield-mcp
export HF_API_KEY=your_api_key
export HF_SECRET=your_secret
3. Get Your API Keys
- Go to https://cloud.higgsfield.ai/api-keys
- Create or copy your API Key and Secret
4. Connect to an MCP Client
Claude Desktop
Edit your Claude Desktop config:
- macOS β
~/Library/Application Support/Claude/claude_desktop_config.json - Windows β
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"higgsfield": {
"command": "npx",
"args": ["-y", "higgsfield-mcp"],
"env": {
"HF_API_KEY": "hf_xxxxxxxxxxxx",
"HF_SECRET": "sk_xxxxxxxxxxxx"
}
}
}
}
Restart Claude Desktop after saving.
Claude Code (CLI)
claude mcp add higgsfield -- npx -y higgsfield-mcp
Then set your environment variables before running Claude Code:
export HF_API_KEY=hf_xxxxxxxxxxxx
export HF_SECRET=sk_xxxxxxxxxxxx
Cursor / VS Code
Add to your .cursor/mcp.json or .vscode/mcp.json:
{
"mcpServers": {
"higgsfield": {
"command": "npx",
"args": ["-y", "higgsfield-mcp"],
"env": {
"HF_API_KEY": "hf_xxxxxxxxxxxx",
"HF_SECRET": "sk_xxxxxxxxxxxx"
}
}
}
}
Tools
Image Generation
| Tool | Model | Notes |
|---|---|---|
generate_image | Soul | Supports style presets and character references |
generate_image_reve | Reve | |
generate_image_seedream | Seedream v4 | Optional camera_fixed param |
edit_image_seedream | Seedream v4 Edit | Edit existing images via prompt |
Video Generation
| Tool | Model | Notes |
|---|---|---|
generate_video | DoP (lite/turbo/standard) | Requires a motion_id |
generate_video_kling | Kling v2.1 Pro | Prompt = camera movement instructions |
generate_video_seedance | Seedance v1 Pro | Prompt = movement/action description |
generate_video_dop_standard | DoP Standard | Optional duration (2β10s) |
generate_talking_head | Speak v2 | Portrait + WAV audio β lip-synced video |
Status & Control
| Tool | Notes |
|---|---|
get_generation_status | Poll Soul/DoP/TalkingHead jobs (job_set_id) |
get_request_status | Poll all other generation jobs (request_id) |
cancel_request | Cancel a queued job |
Characters
| Tool | Notes |
|---|---|
create_character | 1β5 face images β reusable character_id. Costs 40 credits. |
list_characters | All characters with status and thumbnails |
get_character | Single character by ID |
delete_character | Permanently delete a character |
Upload
| Tool | Notes |
|---|---|
upload_image | Upload base64 image β get public URL for use in generation tools |
Lookup
| Tool | Notes |
|---|---|
list_styles | Style presets for generate_image |
list_motions | Motion presets for generate_video |
debug_credentials | Verify keys are loaded |
Resources
Browse these before generating:
higgsfield://stylesβ style IDs forgenerate_imagehiggsfield://motionsβ motion IDs forgenerate_videohiggsfield://charactersβ your character references
How Jobs Work
Tools return either a job_set_id or request_id. Poll the matching status tool until completed, then download the output URL. Results are kept for 7 days.
generate_image(...) β job_set_id β get_generation_status(job_set_id)
generate_image_reve(...) β request_id β get_request_status(request_id)
generate_video_kling(...) β request_id β get_request_status(request_id)
All tools accept an optional webhook_url to receive a callback when the job finishes.
Talking Head Note
Audio must be WAV format. Convert with:
ffmpeg -i speech.mp3 -acodec pcm_s16le -ar 44100 speech.wav
Credits
| Operation | Credits | USD |
|---|---|---|
| Image 720p | 1.5 | $0.09 |
| Image 1080p | 3 | $0.19 |
| Video lite | 2 | $0.13 |
| Video turbo | 6.5 | $0.41 |
| Video standard | 9 | $0.56 |
| Character (one-time) | 40 | $2.50 |
Top up at https://cloud.higgsfield.ai/credits
