Fal.ai
for LuumenAI
Browse fal.ai models, check pricing, and run inference jobs as approved steps
Connect fal.ai and Luumen can search the model catalog, read unit pricing, estimate what a batch will cost, and follow queued requests to completion. When you want work to actually run, it submits the job, uploads input files, or cancels a stuck request — each change previewed and approved first. Auth is a fal.ai API key.
The Fal.ai toolbox
12 tools: 7 read, 5 write. Reads answer instantly. Writes require approval by default. Everything is logged.
- ReadEstimate PricingEstimate pricing for fal.ai model endpoints.
- ReadGet JWKS for Webhook VerificationRetrieve public keys for webhook signature verification.
- ReadGet ModelsDiscover and search fal.ai model endpoints.
- ReadGet Model PricingRetrieve unit pricing for model endpoints.
- ReadGet Queue Request ResultRetrieve the final result of a completed queue request.
- ReadCheck Queue Request StatusCheck the status of a queued request in fal.ai.
- ReadStream Request Status UpdatesStream request status updates via SSE.
- WriteCancel Queue RequestCancel a queued or in-progress request in fal.ai's queue system. Approval by default
- WriteRun Model (Sync)Invoke a fal.ai model synchronously via fal.run, generating new content (images, audio, video, or other model outputs) hosted at fal.ai CDN URLs in the response. Approval by default
- WriteSubmit Async Inference JobSubmit an asynchronous inference job to fal.ai's queue (queue.fal.run). Approval by default
- WriteSubscribe to Async Inference JobSubmit a fal.ai inference job to the async queue and block until it completes (or a deadline is reached), returning the model's final result. Approval by default
- WriteUpload Input FileUpload an input media file (image, audio, or video) to fal.ai's CDN and return a public access_url that can be passed as a model input. Approval by default
One prompt, start to finish
What a governed Fal.ai run looks like inside Luumen.
Questions
Authorize once with API token. Luumen lists the scopes each action needs before you approve the connection, and credentials never appear in the chat.
Read actions answer immediately. Anything that writes — cancel queue request, run model (sync), submit async inference job, subscribe to async inference job, and more — is shown as a plan and requires approval by default, including the 1 action classified as destructive. Administrators configure that per tool, so you decide exactly which actions can ever run unattended.
You decide. Actions are granted per agent, skill, and team, and per environment — production is not staging. Read access can be broad while writes stay narrow.
Every call to Fal.ai — read or write, approved or declined — is recorded with the actor, the input, and the result, and can be linked to the ticket or change record.
Often connected alongside
All integrations
Adobe Marketing Agent MCPAdobe Experience Platform context, sandbox switches, and async task status
AffindaRead parsed documents, collections, and annotations — change them with approval
Agent MailGive the agent its own mailbox for reading threads and sending updates
AI/ML APIInspect assistants, threads, and runs — and change them as approved stepsPut Fal.ai to work with Luumen
Connect in minutes. Every action scoped, approved, and audited from day one.