Your documents, answering questions.
Upload what your team already has. The platform extracts, chunks and embeds it, and answers cite the passage they came from.
- PDF, Office, HTML, text
- Cited by page
- Access per base and group
-
Ingestion without a detour
Extraction runs inside the platform rather than in a third party service, so a document is not sent anywhere to become searchable.
-
Retrieval you can check
Every answer carries numbered sources with their relevance, and opening one shows the passage and its page rather than the whole file.
-
Access is granted, not assumed
A base can be opened to the organization or restricted to groups and people, with viewer, editor and owner roles. The first grant on a base closes it to everybody else.
-
A trail per knowledge base
Uploads, deletions, grants and retrievals are recorded, so the question of who could read what has an answer with a date on it.
What is kept, and for how long
- Retention set per organization, with a legal hold that suspends it
- Documents produced by the platform reaped on their own clock
- Grants included in the access review export
- Deletion removes the chunks and their embeddings, not only the record
What your team gets
Five parts, one account, one set of permissions.
-
Chat and Fusion
Ask across models, grounded in your own documents, with citations you can open.
-
Agents
Automations that do the work, ask before anything irreversible, and leave a trail.
-
Code and Data
Functions, a document store and hosted sites, without leaving the platform.
-
API
An OpenAI compatible endpoint with a budget per key and a receipt per call.
Put one shelf of documents in
A handbook, a set of contracts, or last year's reports. Ask the question you would have asked a colleague.