Quickstart
Create an API key, send your data and follow the training until the model is ready.
1. Create an API key
Open Settings → API keys and create a key. It starts with ts_ and is shown
only once: keep it in a secret manager or an environment variable.
export TYPESTRICT_API_KEY=ts_...Check it:
curl https://api.typestrict.ai/v1/credits \
-H "Authorization: Bearer $TYPESTRICT_API_KEY"2. Send your data
Describe the files and the decision you want. The answer has one upload URL per file, valid for 15 minutes.
curl https://api.typestrict.ai/v1/datasets \
-H "Authorization: Bearer $TYPESTRICT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "support tickets",
"goal": "Which team handles the ticket?",
"files": [{ "filename": "tickets.csv", "content_type": "text/csv", "size": 48213 }]
}'
# { "id": "<dataset>", "uploads": [{ "file_id": "<file>", "filename": "tickets.csv", "url": "https://..." }] }Upload each file to its URL, then confirm it:
curl -X PUT --upload-file tickets.csv "<url>"
curl -X POST https://api.typestrict.ai/v1/datasets/<dataset>/files/<file>/complete \
-H "Authorization: Bearer $TYPESTRICT_API_KEY"3. Follow the training
Every training says which step it is in and what it needs next:
curl https://api.typestrict.ai/v1/trainings \
-H "Authorization: Bearer $TYPESTRICT_API_KEY"
# { "data": [{ "kind": "dataset", "id": "<dataset>", "step": 1, "next_action": "wait", ... }], "next": null }While next_action is wait, we are reading and masking the data. Then the agent structures it and the training moves
on: Trainings lists every next action and the call that answers it.