API
Run your agents from anywhere.
The Cotera API lets you call the agents you build straight from your own code. Trigger a single job, or fan one job out across dozens of agents at once — with the tools and integrations you already configured, and the orchestration handled for you.
More work, in parallel, for less.
Hold the line on cost
Long, serial agent runs keep paying against a large context window the whole time. Split the work across many smaller runs and pick the model that finishes the job for less.
Price per correct answer is the number that matters.Fan out in parallel
One agent per integration page, article, research thread, or code-gen task — running at once. Kick off many runs from your own code instead of working through them one at a time.
See how parallel beats serial.Once, or continuously
Call an agent for a one-off task, or wire it into a recurring workflow that keeps running as new work arrives — same agent, same tools, no rebuild.
Jump to practical examples.One long run is the expensive way to work.
A single cloud or coding agent grinding through a big job runs serially — step after step, all of it paying against one large, high-context window that only grows as the run goes on. The longer it works, the more every token costs. Cotera takes the same job and fans it out across many agents that each stay small, run at once, and finish together.
Pays the premium the whole way
- —Every step bills against a single high- or max-context window that keeps growing.
- —Work happens one item after another, so wall-clock time scales with the backlog.
- —Rate limits, retries, and status tracking are yours to build and babysit.
Cheaper per job, done sooner
- —Each agent stays in a smaller, cheaper context scoped to one piece of work.
- —Runs fire concurrently, so the whole batch finishes in roughly the time of one.
- —Cotera handles orchestration, rate limits, retries, and run status for you.
Integration pages, articles, research, code and doc generation — anything you'd queue up serially becomes a single batch call. You send the work; Cotera manages execution and hands back results.
From prompt to production in three steps.
Build an agent
Create and prompt an agent in Cotera, connecting the tools and integrations it needs to do real work.
Generate an API key
Open Settings, create an API key for your workspace, and authenticate every request with a standard bearer token.
Call the agent
Send one message or a batch of them over HTTPS and get results back — from your own backend, jobs, or workflows.
See the examplesGenerate a key in Settings.
API keys are created and revoked from Settings → API keys in your workspace. Each request carries the key as a standard Authorization: Bearer header.
Pair it with your x-org-id header so calls resolve to the right organization. Keep both server-side — they unlock every agent your workspace can run.
curl https://app.cotera.co/api/v1/resource/agent/{agentId}/chat \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID" \
-H "Content-Type: application/json" \
-d '{
"message": "Enrich this account and update the CRM."
}'A few calls you'll actually make.
Run one agent, fan a batch out across dozens, point an agent at a cheaper model, or check on a run — all over the same REST surface.
Run an agent
POST /api/v1/resource/agent/{id}/chatSend a prompt to an agent and it runs with its configured tools. Add an optional mode to control whether tool calls run automatically.
curl https://app.cotera.co/api/v1/resource/agent/{id}/chat \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID" \
-H "Content-Type: application/json" \
-d '{
"message": "Draft the integration page for Stripe.",
"mode": "auto"
}'Fan out a batch
POST /api/v1/resource/agent/{id}/chat/batchSend a JSON array of up to 50 prompts in a single call. Each one runs as its own agent, in parallel, instead of queuing behind the last.
curl https://app.cotera.co/api/v1/resource/agent/{id}/chat/batch \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID" \
-H "Content-Type: application/json" \
-d '[
{ "message": "Write the Stripe integration page." },
{ "message": "Write the Shopify integration page." },
{ "message": "Write the HubSpot integration page." }
]'Point at a cheaper model
POST /api/v1/resource/agent/{id}/updateUpdate an agent to a different model once, and every run after uses it. Reach for a source-available model like mistral-large-3 or a hosted mini like gpt-5-mini for cheaper fan-out.
curl https://app.cotera.co/api/v1/resource/agent/{id}/update \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID" \
-H "Content-Type: application/json" \
-d '{
"model": "mistral-large-3"
}'Check a run
GET /api/v1/resource/chat/{id}/statusPoll the status of an in-flight or completed run, then pull the full message list when it lands — no websockets or polling logic of your own to build.
# Status of a single run
curl https://app.cotera.co/api/v1/resource/chat/{id}/status \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID"
# Full message list once it completes
curl https://app.cotera.co/api/v1/resource/chat/{id}/list \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-org-id: YOUR_ORG_ID"A few of the endpoints.
/api/v1/resource/agent/{id}/chatSend a prompt to an agent and get back a result./api/v1/resource/agent/{id}/chat/batchRun up to 50 prompts against an agent in a single batch./api/v1/resource/agent/{id}/updateChange an agent — including the model it runs on./api/v1/resource/chat/{id}/statusCheck the status of an in-flight or completed run./api/v1/resource/chat/{id}/listList the messages produced by a run./api/v1/resource/chat/listList the chats and runs across your workspace.WHAT ARE YOU WAITING FOR?
Ready to build?
Stop duct-taping tools.
Stop writing playbooks no one follows.
Start building agents that actually do the work—and do it well.
Cotera: For people who think busywork is boring.
"This is the best product ever when it comes to shipping. Ten on ten recommended. I just can't wait to see what happens with this product."
Rebecca Blount
Director, Customer Success