Most automation stacks charge per API call, which explodes when you need a real-time, visual agent that sees the screen and acts like a human. The Coasty computer use API charges by the agent step. You pay only for vision, session predict, ground, parse, and the actual task runs. This post lists every endpoint, every field, and every cent so you can budget and build reliable, self-driving workflows.
How pricing works
Coasty uses a prepaid USD wallet where 1 credit equals $0.01. You read your API key from COASTY_API_KEY and send requests to https://coasty.ai/v1 with the header X-API-Key or Authorization: Bearer. Each operation is billed at a fixed rate. There are no hidden usage fees beyond credits. Scopes gate access to specific features. Webhooks are HMAC signed with the Coasty-Signature header.
Vision and perception endpoints
- POST /v1/predict costs $0.05. It takes a base64 screenshot, an instruction, and cua_version. Returns actions plus a status. You loop capture, predict, act until status is done.
- POST /v1/sessions costs $0.10. It creates a stateful trajectory memory session. Then POST /v1/sessions/{id}/predict costs $0.04 and yields actions for that session.
- POST /v1/ground costs $0.03. It maps a screenshot plus an element description to x, y coordinates.
- POST /v1/parse is free. It turns pyautogui code into structured actions.
curl https://coasty.ai/v1/predict \
-H 'X-API-Key: $COASTY_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"screenshot": "$(base64 -i screenshot.png)", "instruction": "Click the first button", "cua_version": "v3"}'Task runs drive agents to completion
- POST /v1/runs bills $0.05 per agent step. Use machine_id, task, cua_version (v3 or v4 for autonomous with pass/fail verifier), optional instructions (appended to the base prompt), system_prompt, max_steps, deadline_seconds, on_awaiting_human (pause/fail/cancel), and webhook_url.
- GET /v1/runs lists runs. GET /v1/runs/{id} returns details. POST /v1/runs/{id}/cancel stops a run. POST /v1/runs/{id}/resume continues a paused run.
- GET /v1/runs/{id}/events streams Server-Sent Events with reconnect support using Last-Event-ID. States: queued, running, awaiting_human, succeeded, failed, cancelled, timed_out.
Task runs charge $0.05 per agent step, not per API call.
Workflows and machines
- Workflows are a versioned JSON DSL of runs. POST /v1/workflows creates workflows, POST /v1/workflows/{id}/runs spawns runs from a workflow, and POST /v1/workflows/runs creates ad-hoc inline workflows.
- Workflow step types include task, assert, if, loop, parallel, human_approval, retry, succeed, and fail. Conditions and variables use double-brace inputs.x and stepId.field forms. Hard guards include budget_cents, max_iterations, and deadline_seconds.
- POST /v1/machines provisions a cloud VM. You can start, stop, and snapshot the machine. The agent drives real desktops, browsers, and terminals, not just API calls.
Where this beats brittle automation
Traditional automation relies on brittle CSS selectors, XPath, or API mocks. If a UI changes, your script breaks. The Coasty computer use API gives you a real-time visual agent that sees the screen, follows instructions, and adapts. You can build workflows that handle dynamic layouts, native apps, and unexpected states. The per-step billing means you only pay when the agent is actively driving the UI.
Use this pricing model to build robust computer use agents for testing, data entry, or onboarding workflows. Get a key at https://coasty.ai/developers and start experimenting with the Coasty computer use API.
Want to see this in action?
View Case Studies