How We Use AI

Timace uses AI in a few distinct places. This page discloses what those are, in plain terms, per Article 50 of the EU AI Act. It's a companion to our Privacy Policy, which covers what personal data feeds into these features.

1. The catalog

Facts, taglines, descriptions, and "Best X" / "X vs Y" comparison pages in the Timace catalog are drafted by AI research agents (currently via OpenRouter, calling third-party language models), which crawl publicly available sources and score their own confidence in what they've drafted. Content above a confidence threshold publishes automatically; content below it is queued for a person to review before it goes live. In other words: not everything you read in the catalog has been read by a person first — some of it is AI-drafted and AI-verified, not human-reviewed, at the time it's published. We hold all of it to the same sourcing standard either way, and correct anything found to be wrong.

2. Operator

Operator, the assistant built into the signed-in Timace workspace, is an AI chat feature. It answers using a language model you can choose (including models from Anthropic, OpenAI, and Google), and it's given context from your own workspace — your projects, notes, tasks, mind maps, and recent activity — so it can answer questions about your own work. You're always talking to a model, not a person, when you use Operator.

3. Compose and Build

Compose and Build take a prompt you write and assemble a workflow, stack, or plan document from it, using a language model plus catalog data retrieved for relevance. The result is a draft for you to edit, not a finished recommendation to act on without review.

4. Search

Semantic search uses an embedding model (Voyage AI) to match your query against the catalog by meaning, not just keywords. See our Privacy Policy for how query text is handled.

5. Questions

Email legal@timace.io with anything about how we use AI, or to flag catalog content you believe is wrong.