What file types are supported?

Markdown, plain text (.txt), HTML, DOCX, and PPTX at launch. More formats on the roadmap.

Can I query my buckets from my own app?

Yes. Every enabled bucket is queryable over a REST API, authorized with scoped API keys you create in the dashboard. A key can be limited to specific buckets and can only ask questions, never read or write your files.

How fast is indexing?

Indexing runs automatically in the background. New and changed files are picked up on a regular schedule, so your index stays current without any manual steps.

Can I use my own model or embeddings?

Not during early access, when queries use a model we manage for you. After launch you will be able to bring your own OpenAI, Anthropic, or Cohere keys and any OpenAI-compatible embedding endpoint, including self-hosted models.

What happens when I delete a file from my bucket?

It is automatically removed from the index on the next sync. No manual cleanup needed.

Early access

Turn any bucket into a queryable knowledge base

Auto-index your files and ask questions in plain language, powered by a model we manage for you during early access. Support for your own keys is coming soon.

Join the waitlist

Enable on any bucket
No data migration
Answers cite their sources
Managed model included

Built around your existing stack

No chunking scripts, no vector DB, nothing to configure. Just your bucket, ready to query.

Managed model in early access

During early access, queries run on a model we manage for you. Bring your own OpenAI, Anthropic, or Cohere keys after launch, with no markup from Fil One.

Automatic indexing

Enable a bucket and Fil One keeps its index in sync with your files. Updated files are re-indexed and deleted files are removed, with no manual steps.

Query it your way

Ask questions from the Fil One dashboard or your own app via the REST API and scoped API keys. An MCP server for Claude and Cursor is coming soon.

Custom embedding models

After launch, bring any OpenAI-compatible endpoint, including self-hosted models, and switch providers without re-indexing.

Supported file types

MarkdownPlain textHTMLDOCXPPTX

Up and running in minutes

Connect your bucket

Enable Bucket Intelligence on any existing bucket with one click. No migration or data export needed.

Files get indexed automatically

Fil One extracts, chunks, and embeds your files and keeps the index up to date automatically, using a model we manage for you.

Ask questions, get answers

Query from the dashboard or the REST API. Answers are generated from your indexed content and cite the source files they draw on.

Query it from your own app

Every enabled bucket is a REST endpoint. Authorize requests with scoped API keys, and get back an answer plus the source files it used.

1import requests
2
3res = requests.post(
4    "https://app.fil.one/api/buckets/handbook/query?region=us-east-1",
5    headers={"Authorization": "Bearer sk_rag_your_key"},
6    json={"query": "What is our refund policy?", "top_k": 5},
7)
8data = res.json()
9print(data["answer"])
10print(data["sources"])  # source files the answer cites

Read the API reference

Answers you can trust, data you control

Grounded in your data

Answers are generated only from your indexed files and cite their sources. No outside knowledge, no invented references.

Scoped, revocable keys

Create query-only API keys limited to specific buckets, and revoke them anytime. Keys are stored as hashes, never in plain text.

Resistant to prompt injection

Your files and questions are treated strictly as data, so a malicious document cannot hijack the model or leak other content.

You stay in control

Turn Bucket Intelligence on per bucket, and off just as easily. Only users on your account can query your indexes.

Common questions

Ask your buckets anything

Early access is open. Join the waitlist and be first in line.

Join the waitlist

Free during early access · Requires an active storage plan