Provider REST API

5 actions

Connect Claude to Higgsfield AI

Generate images and video from a prompt, then poll until the file is ready. Toolspoke puts 5 of its actions behind one MCP endpoint that Claude, Cursor, and Codex all speak.

Connect View pricing

Connection: Provider REST API
Authentication: API key
Actions exposed: 5
Cost per call (typical): 1 credit
Adapter: Maintained by Toolspoke

Connected in three steps

1
Install Higgsfield AI
Open the marketplace in your workspace, add Higgsfield AI to the project your agents work in, and it appears on the gateway immediately.
2
Connect the credential
Authenticate with api key. Where to get one, and what it has to be able to reach, is the next section.
3
Point your agent at the gateway
Give your client one address, https://toolspoke.com/mcp. Claude Code takes it as a command, Claude and Claude Desktop add it as a custom connector, and Cursor, Codex and VS Code each read it from a config file of their own.

.mcp.json

{
  "mcpServers": {
    "toolspoke": {
      "type": "http",
      "url": "https://toolspoke.com/mcp"
    }
  }
}

One block covers every tool you have installed. Higgsfield AI shows up in the client as soon as your policy allows it, and so does everything else you install later.

Where the address goes, per client

Claude Code: Run it in your project, then /mcp to sign in
claude mcp add --transport http toolspoke https://toolspoke.com/mcp
Claude and Claude Desktop: Settings, then Connectors, then Add custom connector
https://toolspoke.com/mcp
Cursor: ~/.cursor/mcp.json, or .cursor/mcp.json for one project
{ "mcpServers": { "toolspoke": { "url": "https://toolspoke.com/mcp" } } }
Codex: ~/.codex/config.toml
[mcp_servers.toolspoke] url = "https://toolspoke.com/mcp"
VS Code: .vscode/mcp.json, or the MCP: Add Server command
{ "servers": { "toolspoke": { "type": "http", "url": "https://toolspoke.com/mcp" } } }

What Higgsfield AI asks for

API key. You provide it once, when you install the connector. Toolspoke encrypts it at rest and decrypts it only for the length of a single call, and the gateway attaches it to the outbound request itself, so it is never part of the arguments an agent sends.

API keyRequired: cloud.higgsfield.ai → API keys. Paste the key id and the key secret joined by a colon, in that order - Higgsfield's own header format is `Key {api_key_id}:{api_key_secret}`, and its Python SDK reads the same pair from one HF_KEY variable.; key-id:key-secret

What Claude can do in Higgsfield AI

5 actions, each one declared and named by the connector rather than discovered at runtime. A workspace policy grants a person all of them, a hand-picked selection, everything on the read side, everything on the write side, or none.

Reads: 2Reads
Writes: 3Writes
Destructive: 0Destructive

Reads

Fetches data and changes nothing.

estimate_cost
Price a generation before running it. Send the same model path and the same body `generate` would take, and Higgsfield answers with what it would cost this account in `credits` and in `usd`. Nothing is queued and no credit is spent, which is why this is a read and why it is also the connection test. Always call it before an expensive video generation - cost rises steeply with duration and resolution, and this is the only way to see the number in advance.
get_request_status
Check a generation started by `generate`. `status` is queued, in_progress, completed, failed, nsfw, or canceled - the last four are terminal. Poll about every two seconds at first and back off towards ten. A completed request carries the output: `images` for image models, `video` for video models, `audio` or `audios` for audio. Those URLs are kept for at least seven days and then may be removed, so anything worth keeping has to be copied elsewhere. `failed` and `nsfw` requests are not charged and any reserved credits are refunded.

Writes

Creates or updates something on the other side.

generate
Queue a generation and get back a `request_id`. This spends the account's credits - call estimate_cost with the same arguments first. Nothing but the id and `status: queued` comes back: poll get_request_status until the status is terminal, then read `images`, `video` or `audio` from it. Every model has its own path and its own parameters and refuses parameters belonging to another, so read the model's entry in Higgsfield Cloud before composing `body`. Confirmed paths include higgsfield-ai/soul/standard (text to image), higgsfield-ai/dop/standard (image to video), and kling-video/v2.1/pro/image-to-video. Submissions take no idempotency key, so a call that times out ambiguously must not be blindly repeated - look for the request with get_request_status instead.
cancel_request
Stop a generation that has not started yet, so it stops costing anything. Only works while the request is still `queued` - once it is `in_progress` Higgsfield refuses with a 400 and the work runs to completion. A successful cancellation refunds the reserved credits and returns an empty body. This destroys nothing: there is no output yet to lose.
create_upload_url
Get a place to put an input file that is not already on a public HTTPS URL. Returns `public_url` - the value to pass into `generate` - plus a one-hour `upload_url` and the `upload_headers` the bytes must be sent with. The upload itself is a PUT this gateway does not make: hand `upload_url` and `upload_headers` to whatever holds the file, and never send Higgsfield's credentials to it. Accepted types are image/jpeg, image/png, image/webp, image/gif, audio/wav and video/mp4.

What it will not do

Enforced by the gateway rather than left to convention, which is why each of these can be stated flatly.

It cannot call anything else: The 5 actions above are the whole of it. A call to any other name is refused before it reaches Higgsfield AI rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time.
Nothing here deletes: This connector writes to Higgsfield AI, but nothing in it deletes or permanently alters anything.
It reaches no further than your credential: Toolspoke holds no access to Higgsfield AI of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either.
It never hears from Higgsfield AI: Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Higgsfield AI. An agent has to ask.
It does not smooth over provider limits: Toolspoke does not retry, queue or back off around Higgsfield AI's own rate limits. A call that Higgsfield AI refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in Higgsfield AI?

5 named actions: 2 that only read and 3 that write. They include estimate_cost, generate and get_request_status. Nothing outside that list is reachable: the connector declares each operation by name rather than proxying whatever an agent asks for.

What credentials does the Higgsfield AI connector need?

API key. The connector asks for api key. Values are encrypted at rest and attached to the outbound request by the gateway, so they are never part of the arguments an agent sends and never reach the audit log.

Does the Higgsfield AI connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 5 actions. There is a single address, https://toolspoke.com/mcp. Claude Code adds it with claude mcp add --transport http, Claude and Claude Desktop take it as a custom connector in settings, Cursor reads it from .cursor/mcp.json, Codex from ~/.codex/config.toml, and VS Code from .vscode/mcp.json. Each of them signs in to the gateway itself, so there is no key to paste.

What does the Higgsfield AI connector not do?

The 5 actions above are the whole of it. A call to any other name is refused before it reaches Higgsfield AI rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time. This connector writes to Higgsfield AI, but nothing in it deletes or permanently alters anything. Toolspoke holds no access to Higgsfield AI of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either. Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Higgsfield AI. An agent has to ask. Toolspoke does not retry, queue or back off around Higgsfield AI's own rate limits. A call that Higgsfield AI refuses comes back to the agent as a failed call.

Can I limit which actions an agent can call?

Yes, in two places. The project switches Higgsfield AI's actions on and off one at a time, for everyone in the project at once, and the screen groups them by read, write and destructive so turning off everything that deletes is one click. An individual agent key can then be narrowed further, to particular toolkits in a project and to particular actions in a toolkit. Whatever it was granted, a key never reaches a project its owner cannot.

What gets recorded when an agent calls Higgsfield AI?

Every attempt, with the agent that made it and the person that agent belongs to, the full request payload, the response payload, the status, the duration, and the credits spent. Values whose key names a secret are masked out before the record is shown to anyone. An operation the connector marks as not retained never has its response body written at all, so the gateway keeps no second copy of what was read.

More ai connectors

Browse all of them