1Onboarding Policy.pdfp. 2
Onboarding is role-based, with a two-week check-in for every new hire.
01Add a document
02Ask
What changed in the onboarding policy?
03Inspect
Grounded answer
The policy moves onboarding to a Onboarding Policy.pdf · p. 2Onboarding is role-based, with a two-week check-in for every new hire. path, adds a Onboarding Policy.pdf · p. 2Onboarding is role-based, with a two-week check-in for every new hire., and makes Security Checklist.md · p. 1All new hires must complete security training in the first week as part of the checklist. part of the first-week checklist.
Open source
2Security Checklist.mdp. 1
All new hires must complete security training in the first week as part of the checklist.
SOURCES
Browser-local workspace
Browser-local index
Citations you can inspect
Visible pipeline
Independent comparisons
Easy by design
From file to evidence in three clear steps.
Follow one question from two documents to a cited answer. Scroll through the steps, then open the references yourself.
- 1 / 3
Add what you already have
Drop PDFs, DOCX, notes, code, or a public page. RAG Studio turns readable text into an inspectable local index.
Try this in StudioRAG StudioIllustrative example · not your workspace dataAdd a document
Onboarding Policy.pdf · p. 2
Onboarding is role-based, with a two-week check-in for every new hire.
Security Checklist.md · p. 1
All new hires must complete security training in the first week as part of the checklist.
- 2 / 3
Ask in plain language
Start with the question you actually need answered. Retrieval finds the passages before generation begins.
Try this in StudioRAG StudioIllustrative example · not your workspace dataRetrieved evidence
What changed in the onboarding policy?
[1] Onboarding Policy.pdf · p. 2
Onboarding is role-based, with a two-week check-in for every new hire.
[2] Security Checklist.md · p. 1
All new hires must complete security training in the first week as part of the checklist.
- 3 / 3
Check the evidence
Open citations, compare retrieval choices, and inspect the source behind each answer before you trust it.
Try this in StudioRAG StudioIllustrative example · not your workspace dataGrounded answer
The policy moves onboarding to a role-based path, adds a two-week check-in, and makes security training part of the first-week checklist.
Hover, focus or tap a referenceAnswer linked to sources
Nothing hidden
Follow the answer all the way back.
The same pipeline overview that powers the Studio becomes a simple map: sources become passages, retrieval selects evidence, the answer cites it, and evaluations show what to improve.
PreviewIllustrative example · not your workspace data
Stage inspection
01 / 05Start with readable text
Each source keeps its identity. Here, two example documents provide the evidence for one onboarding question.
Onboarding Policy.pdf · p. 2
01Onboarding is role-based, with a two-week check-in for every new hire.
Security Checklist.md · p. 1
02All new hires must complete security training in the first week as part of the checklist.
Local-first economics
Keep the core small.
Spend only when it helps.
The current workflow parses, chunks, indexes, retrieves, and inspects evidence in your browser. Add a hosted provider—or a reachable local model—only when you want generated answers.
See your optionsLocal workspace
Documents, chunks, conversations, and evaluation sets stay in this browser profile.
No required vector service
Try the complete inspection workflow before adding hosted infrastructure.
Inspect before scaling
See chunk boundaries and retrieval results before spending on a larger stack.
Bring your own model
Use a provider session key or bind a reachable local runtime when generation matters.
Commercial paths
Use it online.
Or keep it inside your company.
RAG Studio is local-first by default. Online subscriptions, usage wallets, and managed team access are being prepared; private deployment and setup can be discussed now.
Early access
Online workspace
Use the hosted Studio immediately with browser-local documents. Bring your own provider key today; managed model credit is planned.
- One personal workspace
- Browser-local corpus
- Your own model key
Planned
Team subscription
A seat-based plan for teams that need shared policy, predictable limits, and one place to manage model usage.
- Seats and shared workspaces
- Usage caps
- Central billing
Available by arrangement
Private deployment
Run RAG Studio on your own server or private cloud, with installation, model connection, and onboarding handled with you.
- Your infrastructure
- Local or hosted models
- Setup and support
Pay only for the model usage you need
The planned wallet separates the Studio subscription from provider cost: add credit, set a limit, and spend it only when a hosted model runs.
Support independent development
If RAG Studio helps your work, you can sponsor the project or fund a specific capability without buying a team deployment.
Become a sponsorContact for a private setup
Before you start
Questions, answered plainly.
01Do my documents leave the browser?+
The current workspace stores documents and its index in this browser profile. If you choose a hosted model and run generation, the current question, recent conversation context, and retrieved passages are sent with your session key.
02Do I need an API key to try it?+
No. Document parsing, chunking, local retrieval, citations, graphs, and the extractive fallback work without a provider key. Provider-generated answers need your own key or a reachable local runtime.
03Which source types are supported?+
PDF, DOCX, Markdown, plain text, HTML, common code files, pasted text, and safe public web pages are supported in the current browser workflow.
04Is multilingual retrieval certified?+
Not yet. Unicode text can be ingested, but retrieval quality depends on the language and model. We intentionally do not show a multilingual badge until those paths have language-specific tests.