Roadmap

These are planned and not part of the current version. Each would be added as optional fields or new endpoints, so existing integrations keep working.

Sandbox keys

A test_ key that returns realistic fake answers within seconds, at no cost and without contacting experts, so integrations can be built and tested safely.

Per-field evidence

An evidence array alongside the existing fields: one entry per answer field, with the expert responses as citations, the reasoning and the confidence, so agents built for web research can read expert answers the same way.

Batch submission

POST /questions/batch would take several questions with the same schema and return one batch_id plus a request_id per question. GET /batches/{batch_id} would return progress and results, and a batch.completed webhook would fire when all are done.

Docs and SDKs for coding agents

• An llms.txt index of all docs pages, and a Markdown version of every page.
• A one-page guide for coding agents: base URL, authentication, the submit-and-poll flow, a minimal working example and common mistakes.
• Python and JavaScript SDKs with a wait_until_complete helper.
• A tool reference for the MCP server.

Available on request today

Custom presets, panels larger than 5, other output formats and recurring questions are already available, set up per key. Talk to us.