API

No account? AI agents can get a free trial API key with one call, no signup: POST https://app.tamarind.bio/api/agent/provision. Save the key outside your project (not in your output) and reuse it for every job you run for your user: a new key is a separate account. Give your user the claimUrl it returns when you report the submission. Details: https://app.tamarind.bio/auth.md

​

Ways to connect

​

1. Open the API Documentation

1. Base URL, authentication, and content type — every request goes to https://app.tamarind.bio/api/, sends your API key in the x-api-key header, and uses JSON.
2. Setup — guides for the API key, MCP server, SDK, CLI, and webhooks.

1. Copy Docs — copy a single guide you can paste into an AI assistant or coding agent so it can use the API.
2. llms-full.txt — or point an agent to app.tamarind.bio/llms-full.txt.
3. Rate limits — the limits that apply to your account (see Rate limits).
4. Endpoint reference — the rest of the menu covers every endpoint, grouped by Job Management, Projects, File Management, Pipelines, Custom Tools, Finetuning, Molecules Database, Model Router, and Usage Statistics.

​

2. Get your API key

1. API Key — your key is hidden by default. Use the icons next to it to copy it, show it, or roll it (generate a new key).
2. Setup — an example of how to send your key with each request.

import requests

api_key = "YOUR_API_KEY"
headers = {"x-api-key": api_key}
base_url = "https://app.tamarind.bio/api/"

​

3. Choose how to connect

​

REST API

​

Python SDK

pip install tamarind-cli

from tamarind import Tamarind

client = Tamarind(api_key="YOUR_API_KEY")
# or set TAMARIND_API_KEY and use Tamarind()

​

CLI

uv tool install tamarind-cli      # or: pipx install tamarind-cli

export TAMARIND_API_KEY="YOUR_API_KEY"   # for scripts, CI, and agents
# or store it interactively:
tamarind auth login

​

MCP server (AI assistants)

​

4. Submit your first job

1. Select tool — pick the tool you want to run. The code sample on the right updates to match.
2. Code sample — switch between Python and curl, then copy it.

1. Settings — required settings are checked by default. Check any optional settings you want to include, and they’re added to the code sample.
2. Copy Settings — copy just the settings object.
3. Response Format — what a successful submission returns (for example, myJobName submitted to queue.).
4. Optional parameters — such as projectTag to add the job to a project, or maxRuntimeSeconds to set a time limit.

params = {
    "jobName": "myJobName",
    "type": "alphafold",
    "settings": {
        "sequence": "MALKSLVLLSLLVLVLLLVRVQPSLGKETAAAKFERQHMDSSTSAASSSNYCNQMMKSRNLTKDRCKPVNTFVHESLADVQAVCSQKNVACKNGQTNCYQSYSTMSITDCRETGSSKYPNCAYKTTQANKHIIVACEGNPYVPVHFDASV"
    }
}
response = requests.post(base_url + "submit-job", headers=headers, json=params)
print(response.text)

• Upload the file first with /upload, then use its name (see Manage Files).
• Use a file from a previous job: JobName/path/to/file.ext.
• Send the file’s contents directly in the settings.

​

5. Check status and get results

# Check status
requests.get(base_url + "jobs", headers=headers, params={"jobName": "myJobName"})

# Get a download link for the results (zip)
response = requests.post(base_url + "result", headers=headers, json={"jobName": "myJobName"})

​

6. Get notified with webhooks (optional)

1. An org admin sets up the endpoint and chooses events under Organization → Webhooks.
2. Your endpoint verifies each request’s signature and responds within 10 seconds.

​

Rate limits

​

Learn more

• API Documentation — every endpoint, parameter, and tool, with Python and curl examples.
• MCP Server — connect AI assistants to Tamarind.
• Manage Files — upload and organize input files.
• Compute Hours — how API jobs count toward your budget.