Higgsfield MCP server for AI agents
Let Claude, ChatGPT or your own agent read Higgsfield: 5 typed functions through one MCP server. No Higgsfield API key needed.
- Site
- higgsfield.ai
- Functions
- 5
- Sign-in
- Not needed
- Works with
- MCP, HTTP, Node, Python
Connect
An MCP server for Higgsfield alone, with one tool per function. Paste it into Claude, ChatGPT, Cursor or any MCP client as a remote server.
https://api.bowmark.ai/mcp/site/higgsfieldOne MCP server for Higgsfield and every other site Bowmark covers. Your agent looks up what it needs and calls it.
https://api.bowmark.ai/mcpSet it up in your clientBoth need a Bowmark API key. The quickstart gets you one and a first result in a few minutes.
Functions
Open a function to see its arguments, what it returns and how to call it.
estimateCostAsks Higgsfield what one generation will cost, for exactly the parameters that would be submitted, without submitting it.Arguments
Name Type Required args{ prompt: string; model?: string; [param: string]: unknown }Yes Returns
higgsfieldEstimateinterface higgsfieldEstimate { model: string; credits: number; usd: number; discountUsd: number | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/higgsfield/estimateCostCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.higgsfield.estimateCost({ model: "soul", prompt: "a quiet alpine lake at sunrise" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.higgsfield.estimateCost({ "model": "soul", "prompt": "a quiet alpine lake at sunrise" }))curl -X POST https://api.bowmark.ai/v1/sites/higgsfield/estimateCost \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"model":"soul","prompt":"a quiet alpine lake at sunrise"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.higgsfield.estimateCost({ model: "soul", prompt: "a quiet alpine lake at sunrise" });generateImageTurns a text prompt into finished images and waits for them, returning their URLs plus what Higgsfield quoted.Arguments
Name Type Required argsstring | { prompt: string; model?: string; waitMs?: number; save?: boolean; [param: string]: unknown }Yes Returns
higgsfieldGenerationinterface higgsfieldEstimate { model: string; credits: number; usd: number; discountUsd: number | null; } interface higgsfieldRequest { requestId: string; status: "queued" | "in_progress" | "completed" | "failed" | "nsfw" | "canceled"; imageUrls: string[]; videoUrl: string | null; error: string | null; statusUrl: string | null; cancelUrl: string | null; } interface higgsfieldGeneration extends higgsfieldRequest { model: string; estimate: higgsfieldEstimate; settled: boolean; files: higgsfieldSavedFile[]; // the same media, kept in YOUR account, Higgsfield's own links expire warnings: string[]; } // Bowmark's own saved-file handle (`SavedFile`), spelled out here because the // published surface has to stand alone. Manage these with `bowmark.files.*`; // `url` is presigned and `bowmark.files.url(id)` mints a fresh one. interface higgsfieldSavedFile { id: string; name: string; contentType: string; bytes: number; url: string; expiresAt: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/higgsfield/generateImageCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.higgsfield.generateImage({ model: "soul", prompt: "a quiet alpine lake at sunrise", aspect_ratio: "99:1" });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.higgsfield.generateImage({ "model": "soul", "prompt": "a quiet alpine lake at sunrise", "aspect_ratio": "99:1" }))curl -X POST https://api.bowmark.ai/v1/sites/higgsfield/generateImage \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"model":"soul","prompt":"a quiet alpine lake at sunrise","aspect_ratio":"99:1"}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.higgsfield.generateImage({ model: "soul", prompt: "a quiet alpine lake at sunrise", aspect_ratio: "99:1" });generateVideoTurns a text prompt, or a prompt plus an input image, into a finished video and waits for it, returning its URL plus what Higgsfield quoted.Arguments
Name Type Required argsstring | { prompt: string; model?: string; waitMs?: number; save?: boolean; [param: string]: unknown }Yes Returns
higgsfieldGenerationinterface higgsfieldEstimate { model: string; credits: number; usd: number; discountUsd: number | null; } interface higgsfieldRequest { requestId: string; status: "queued" | "in_progress" | "completed" | "failed" | "nsfw" | "canceled"; imageUrls: string[]; videoUrl: string | null; error: string | null; statusUrl: string | null; cancelUrl: string | null; } interface higgsfieldGeneration extends higgsfieldRequest { model: string; estimate: higgsfieldEstimate; settled: boolean; files: higgsfieldSavedFile[]; // the same media, kept in YOUR account, Higgsfield's own links expire warnings: string[]; } // Bowmark's own saved-file handle (`SavedFile`), spelled out here because the // published surface has to stand alone. Manage these with `bowmark.files.*`; // `url` is presigned and `bowmark.files.url(id)` mints a fresh one. interface higgsfieldSavedFile { id: string; name: string; contentType: string; bytes: number; url: string; expiresAt: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/higgsfield/generateVideoCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.higgsfield.generateVideo({ model: "hailuo-02-standard", prompt: "a heron landing on still water", duration: 999 });import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.higgsfield.generateVideo({ "model": "hailuo-02-standard", "prompt": "a heron landing on still water", "duration": 999 }))curl -X POST https://api.bowmark.ai/v1/sites/higgsfield/generateVideo \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":{"model":"hailuo-02-standard","prompt":"a heron landing on still water","duration":999}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.higgsfield.generateVideo({ model: "hailuo-02-standard", prompt: "a heron landing on still water", duration: 999 });getRequestStatusChecks one submitted generation by the request id a generate call returned, and hands back its current state plus any finished media.Arguments
Name Type Required argsstring | { requestId: string }Yes Returns
higgsfieldRequestinterface higgsfieldRequest { requestId: string; status: "queued" | "in_progress" | "completed" | "failed" | "nsfw" | "canceled"; imageUrls: string[]; videoUrl: string | null; error: string | null; statusUrl: string | null; cancelUrl: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/higgsfield/getRequestStatusCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.higgsfield.getRequestStatus("1c96cfb4-1ecb-452f-828c-47d0fa0dae6c");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.higgsfield.getRequestStatus("1c96cfb4-1ecb-452f-828c-47d0fa0dae6c"))curl -X POST https://api.bowmark.ai/v1/sites/higgsfield/getRequestStatus \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"args":"1c96cfb4-1ecb-452f-828c-47d0fa0dae6c"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.higgsfield.getRequestStatus("1c96cfb4-1ecb-452f-828c-47d0fa0dae6c");listModelsLists every generation model Bowmark's Higgsfield account can actually call.Arguments
None.
Returns
higgsfieldModel[]interface higgsfieldModel { id: string; path: string; kind: "image" | "video"; needsInputImage: boolean; family: string; measuredUsd: number; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/higgsfield/listModelsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.higgsfield.listModels();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.higgsfield.listModels())curl -X POST https://api.bowmark.ai/v1/sites/higgsfield/listModels \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.higgsfield.listModels();
Use it from code
- Node
npm i @bowmark/web - Python
pip install bowmark-web - HTTP
POST /v1/sites/higgsfield/<function> - OpenAPI
/v1/sites/higgsfield/openapi.json