The engagement
We build it, we run it. Two ways to buy.
Custom workflows get built next to the solutions we already run, which is why they land fast and cost less than starting from zero. Anything already live in the catalog below carries no build fee.
Implementation
One-timeWe scope the workflow, build it, host it, and hand you one endpoint. API calls are billed per call at list rates, and upkeep is billed monthly as models change. On the commitment, both come included.
- A dedicated team scopes and builds the workflow with you
- Built on the Solutions we already run in production, which is why it costs less than starting from zero
- Every model it touches runs on our infrastructure, from 700+ open-source and frontier models
- Sentinel scores every output before it reaches your users
- Example build: the 3D-to-AI image pipeline we shipped for a client
Scoped on a 25-minute call. You get a fixed price before any work starts.
Managed
12-month termA monthly minimum in API spend on a 12-month term. The commitment buys usage, and it unlocks the builds, the maintenance, and lower locked rates.
- The minimum is API spend, not a service fee: what you commit, you consume in usage
- Workflow builds included, earned through use
- Maintenance on us: when a new model lands we update, re-benchmark, and keep it current
- Per-call rates lock lower than list, and volume discounts stack as your spend grows
- We run the pipeline in production and carry the on-call
- Sentinel installed inside your company, managed by us, operated by your team
Same call. We tell you which of the two paths fits your volume.
Out of the box
18 solutions ready on day one
Every engagement starts from these pipelines, already running in production. They carry no build fee: book a demo, we wire them into your product, and you pay the per-image price. Committed volume gets its own rate, agreed in the contract.
Per image, simple fixed pricing. Details on each solution page, and every model page publishes its per-call price.
The engagement
What we ship, and what we keep running
One engagement covers the whole path from a scoping call to a production endpoint your team calls every day. We staff it with our own people and deliver it with our own software, which is why it lands at roughly 70% cheaper vs in-house. BetterPic ran this motion first, on 35M+ images.
What our team does
- We scope the pipeline around your inputs, outputs, and quality bar
- We build it next to the Solutions already running in production
- We host every model it needs, open source and frontier. You never provision a GPU.
- We run it, score every output with Sentinel, and send one invoice
3D-to-AI image pipeline
We built a pipeline that turns raw 3D model renders into finished, photoreal product imagery, then scores every frame before delivery. One endpoint for the customer, a full custom workflow underneath.
FAQ