Free
Try the full value
Free
- 1,000 Active corpus/month
- 1,500 Simple Retrieval/month
- 5 Agentic RAG/month
- Other limits: Workspaces: 5 · Areas: 4 · Collections: 5 · Synthesis (source pages): 1,500 · Assists: 10
RAG service
Your AI, answering from your documents.
Augmented retrieval for your AI.
RAGfly searches all your documents and hands your AI only the paragraphs that answer, with their source.
Simple, private, and precise. Access by workspace, area, role and agent.
Why a RAG
Your AI can answer from one document, or a few. From hundreds or thousands, it can't.
It would have to read them all, or you would have to tell it which one to open.
That's what a RAG is for. It quickly searches all your documents and hands your AI only the paragraphs that answer the question.
- Answers with what the AI doesn't know: your private documents.
- Cites the source, so you can check.
- Doesn't pay to read 16 pages when one paragraph was enough.
HOW IT WORKS
From your folders to answers with sources.
How do I stop the ACME contract from renewing? You have two contracts with ACME: Service agreement.pdf: Give written notice 60 days before it expires. Support agreement.pdf: Email notice 30 days before it expires. I found 32 other ACME documents on contract terms.
Choose your folders.
With RAGfly Desktop, on Mac or Windows. RAGfly reads, organizes and indexes them on its own — scans included.
We don't store your documents.
Try it in the chat.
Ask RAGfly's chat about your documents and see how it answers. Every answer comes with its source: the document and the page.
Connect your AI.
Then put it to work: Claude, ChatGPT or your own agents query your documents whenever they need them.
Over MCP, REST, the SDK or the CLI. · Connection guide →
PREFLIGHT · THE REFUSAL
Verified answers, not plausible ones.
A verifier breaks the answer into claims and checks each one against its evidence before showing it. If something isn't backed, it doesn't ship.
You can only trust a system that knows how to say no.
PREFLIGHT
- corpus indexedPENDING
- permissions appliedPENDING
- citation resolved to document and pagePENDING
- evidence for this claimevidence for this claimPENDINGNO EVIDENCE
SIMPLE · BOARDING
You point at the directory. The pipeline does the rest.
You bring the directory —your local folder— and the system loads, extracts, analyzes, chunks and vectorizes on its own. No document ingestion to build, no upfront configuration and no integration project.
Google Drive and your documents
When you connect Google Drive, you authorize RAGfly to read documents in your selected folder and subfolders. We extract their text to generate vector representations and a RAG index, enabling your AI to retrieve relevant information with its sources. You can combine this knowledge with documents from other sources and control access through workspaces, users and profiles. RAGfly requests read-only access and does not modify or delete your Drive files.
Privacy PolicyPIPELINE STATES
- LOADED
- METADATA
- SCANNED
- CHUNKED
- VECTORIZED
The climb is smooth; the cruise, instant.
Ingesting a large corpus runs on its own, unattended: you leave it going and move on. What you notice in every query isn't speed — it's that the answer is faithful to what you loaded.
Onboarding without red tape
If you start out alone, account, workspace, area and profile are a single node that unfolds as your organization grows. The simplicity of day one, with the enterprise ceiling already built in.
INPUT · FORMATS
PDF, Word, Excel, PowerPoint, RTF, plain text and scans via OCR.
What we don't doWe don't process audio, video, or non-text images.
OUTPUT · INTERFACES
Web chat, MCP server, REST API /v1, Python SDK, TypeScript SDK and CLI.
On every planAll six, from a single configuration. RAGfly Desktop is included on every plan, Free included.
PRIVATE · THE MANIFESTO
We don't store your documents. And if you want, they never leave your network.
In Cloud the original is processed on the way in and not retained: what persists are the derivatives —chunks, embeddings and metadata—. With Desktop, the file and its extraction stay inside your network, and depending on the deployment derivatives may be synchronized. On sovereign plans, that database is yours.
ISOLATION
It's not that you can't see it. It's that it isn't on board.
The workspace filter lives inside the search engine itself: vectors from another workspace are never even scored. It isn't a filter applied afterwards — they never entered the search. It holds on every plan, the free one included.
With Desktop, derivatives are synchronized, not the original file.
Depending on the deployment, chunks, embeddings and metadata may be synchronized. RAGfly processes the chunks needed to retrieve context.
MI-PASS
Workspace → Area → Role → Agent
The workspace is the line you draw: your personal life and your business, or one client per workspace. Nothing crosses it. Inside, the area defines which document subtree each identity sees — with hierarchical inheritance: whoever sits above sees what their descendant areas hold —, the role defines what it can do, and the user or agent comes in with its own permissions.
With no area assigned, access is closed by default, never open.
The text of a forbidden document never reaches the model or the agent: it isn't hidden on screen, it doesn't enter the prompt.
PRIVACY
A service guarantee, not a checkbox.
On every plan, RAGfly selects providers whose terms exclude training on your data. If you choose or supply your own providers, their policies apply.
INSTRUMENTS
AES-256-GCM · application-level encryption with key hierarchyTLS 1.3 in transitAudit log of business operations
PRECISE · THE TRAIL
Retrieves by meaning and shows where it came from.
The answer doesn't end at the text: it ends at the exact chunk that backs it, with its page.
CITATIONS
Direct link to the document and the page
Every claim carries its citation. The trail runs from the answer to the chunk that backs it, and you can follow it.
How we get there
RETRIEVAL
Hybrid search + rerank
Three branches —vector search over pgvector, lexical search over what the system understood from the document, and a dedicated branch for the summary— fused and reordered by a multilingual reranker.
If rerank isn't available, the degradation is visible, not silent.
EVALUATION
Precision measured, not declared
An evaluation suite with verifiable ground truth: the question is generated from the real chunk that answers it, so the correct answer is known in advance.
TRIAGE BY LEVEL
- 01 · retrieval
- 02 · generation
- 03 · citation
- 04 · numeric accuracy
Improvement is targeted, not guesswork.
PROFILES · THE CREW
You don't give your agent access to the documents. You give it a profile.
The agent signs in as «Finance Analyst» and sees exactly what that profile would see — under the same access rules as your people.
Service identity
01A profile is an identity with no personal data, designed for agents. It inherits the same machinery as a person: area with hierarchical inheritance, associated locations, roles, RBAC and identity prompts. No master keys.
Frozen scope
02Workspace, area and role are pinned: not even the agent itself can ask for a different scope.
Credential issued to the identity
03The credential is issued in the name of an identity and inherits its scope: change the profile's area and what the agent sees changes, without touching the integration.
Scales with your portfolio
04Twenty workspaces are twenty isolated sets of profiles, governed from a single place. You define the profile once and assign it to any agent, today or next year.
In a regulated organization, it guarantees an agent never leaks anyone a document they weren't supposed to see.
CONFIGURATION
Configuration expands the context.
Everything you configure becomes context your AI receives — editable from the screen, with no deploy and no developer. It's what separates RAGfly from a vector database with a chat on top.
SYSTEM PROMPTS01
Seven layers, five editable
Product → application → function → your business hierarchy (workspace, area, role, profile) → capability catalog. You write «this workspace is a construction company; its documents are usually works contracts» and from the next message on, everyone answers through that lens.
A change in a text box, new behavior instantly.
TYPOLOGY02
Types and attributes
The pipeline doesn't chunk blindly: it types each document and extracts only the attributes that apply to its type —amount, counterparty, due date—.
That's what enables exact answers: «how many invoices from May?», «the contract with the highest amount».
ENRICHMENT03
Context in every vector
Before vectorizing, each chunk is prefixed with the whole document's record: name, location, date, summary and attributes. Enrichment, not just chunking.
It's measured: we tried removing it and retrieval got worse.
The summary also competes as a candidate of its own in the search.
FOR DEVELOPERS
Use it from wherever you work. Claude, ChatGPT… over MCP, REST, CLI or SDK.
All six interfaces —web chat, MCP, REST, Python SDK, TypeScript SDK and CLI— are on every plan, the free one included. RAGfly Desktop too.
Your AI (over MCP)
Start hereConnect RAGfly as an MCP server and your AI converses with the documents without you building any UI. It's where we open the door.
CLI
DX, operations and bulk ingestion. Day-to-day convenience.
REST
The universal production foundation. The base everything else sits on.
SDK
Integrate RAGfly directly into your code. Python and TypeScript. No MCP infrastructure required.
Connect your AI (over MCP)No integration project.
{
"mcpServers": {
"ragfly": {
"type": "http",
"url": "https://api.ragfly.ai/mcp-http",
"headers": { "Authorization": "Bearer ${RAGFLY_API_KEY}" }
}
}
}NOT JUST CONTEXT
Not just context: action
The agent doesn't stop at retrieving chunks: it invokes operations over sets of documents by prompt, from your own agent.
«Download these 500 documents and put them where I tell you.»
RAGfly is a toolset of document operations that the agent calls, not just a context provider.
YOUR CREDENTIALS
With parsing-BYOK and embedding-BYOK, no model under our account ever sees your document.
The five pipeline calls —reading, typing, summary, attributes and vectorization— run on your credentials, against Anthropic, OpenAI or Google, under your contract and your invoice.
We never retry with ours.
If your credential fails —quota, rate limit, provider down—, the operation fails and we tell you. A silent fallback would break exactly that promise.
NO LOCK-IN
Your models, your database, your documents.
You're always in control.
MODELOWNER · YOU
Your models
With parsing-BYOK and embedding-BYOK, processing runs on your credentials, against Anthropic, OpenAI or Google, under your contract and your invoice.
DATABASEOWNER · YOU
Your database
Vectors can live in your own database instead of ours: on Scale, your own Supabase —the same engine that runs RAGfly, Postgres with pgvector—; on Enterprise, other engines.
DOCUMENTSOWNER · YOU
Your documents
You never lose control of your documents: we never store them.
With parsing-BYOK and embedding-BYOK, content is processed using your credentials. On sovereign plans, vectors are stored in your own database.
Plans
Clear plans for simple, private and precise RAG.
PRICES & QUOTAS · USD/MONTH · LIVE SOURCEStarter
Start in production
$39/month
- 10,000 Active corpus/month
- 5,000 Simple Retrieval/month
- 50 Agentic RAG/month
- Other limits: Workspaces: 25 · Areas: 12 · Collections: 10 · Synthesis (source pages): 10,000 · Assists: 30
- Add-ons: corpus +10,000 for $10/month · Agentic RAG +50 for $59.
Recommended
Growth
For teams needing control and operations
$149/month
- 50,000 Active corpus/month
- 25,000 Simple Retrieval/month
- 150 Agentic RAG/month
- Other limits: Workspaces: 150 · Areas: 50 · Collections: 50 · Synthesis (source pages): 50,000 · Assists: 100
- Add-ons: corpus +10,000 for $10/month · Agentic RAG +50 for $59.
Scale
For production at scale
$590/month
- 250,000 Active corpus/month
- 100,000 Simple Retrieval/month
- 500 Agentic RAG/month
- Other limits: Workspaces: 500 · Areas: 200 · Collections: 200 · Synthesis (source pages): 200,000 · Assists: 250
- Add-ons: corpus +10,000 for $10/month · Agentic RAG +50 for $59.
Business
For high volume at a fixed price
$1,990/month
- 1,000,000 Active corpus/month
- 1,000,000 Simple Retrieval/month
- 1,250 Agentic RAG/month
- Other limits: Workspaces: 1,500 · Areas: 1,000 · Collections: 1,000 · Synthesis (source pages): 750,000 · Assists: 1,000
- Add-ons: corpus +10,000 for $10/month · Agentic RAG +50 for $59.
Enterprise / Sovereign
Sovereignty and custom terms
Custom
- Custom Active corpus
- Custom Simple Retrieval
- Custom Agentic RAG
- Other limits: Workspaces: Custom · Areas: Custom · Collections: Custom · Synthesis (source pages): Custom · Assists: Custom
Free is permanent, needs no card and has monthly allowances. Paid plans support manual packs and recurring corpus capacity; there is no automatic overage. Included processing is subject to reasonable use to protect shared performance.
CLOSING
Point RAGfly at your documents. Try it in the chat, then connect your AI.
FREQUENTLY ASKED QUESTIONS
What people ask us before starting
What is RAGfly?
A service that helps your AI answer from your documents. You choose your folders; RAGfly reads, organizes and indexes them — scans included — so your AI, like Claude or ChatGPT, can query them. You can try it first in RAGfly's chat. Every answer comes with its source, and each workspace stays isolated from the rest.
Who is it for?
For people and teams with lots of documents who want to use them with AI: professionals, companies and consultancies — and developers building their own agents.
Do I need to build an agent?
No. Connect the AI you already use, like Claude or ChatGPT, over MCP — or use REST, the SDK or the CLI. Your AI can help with setup. Before that, you can try RAGfly in its own chat, with your documents.
How does it work?
You choose your folders with RAGfly Desktop, on any plan, and RAGfly scans, indexes and vectorizes them on its own. You can try it in its chat; then your AI asks over MCP, REST or the CLI and gets answers with citations to the source, isolated by workspace.
Why does RAG answer better than searching your files?
Because it works with fragments, not whole files. A file system searches names, paths or exact text and hands back entire files that your AI has to read in full. A RAG splits your documents into fragments (chunks) and searches them by meaning: it understands questions in natural language, finds the relevant fragments very quickly, and gives your AI only what it needs, with the source.
How much does it cost?
Plans in USD: Free $0, Starter $39/month, Growth $149/month, Scale $590/month, Business $1,990/month and Enterprise by agreement. Each plan shows active corpus, Simple Retrieval, Agentic RAG and workspaces; paid plans support manual add-ons and have no automatic overage. MCP, REST, CLI and SDK are available on every plan.
What's the difference between Free and paid plans?
Limits and capacity. Privacy is the same on every plan. Free is permanent, with monthly limits, and needs no credit card.
Who processes my documents?
Models from providers whose terms exclude training on your data, on every plan, Free included. RAGfly doesn't use your documents to train models of its own.
What does RAGfly process, and where does my data end up?
RAGfly processes document information on the way in and relevant chunks on the way out. In Cloud it processes the original file but does not retain it after ingestion; chunks, embeddings and metadata persist. With Desktop, the original and its extraction are processed inside your network, although derivatives may be synchronized. On sovereign plans, the database and the derivatives at rest belong to the customer; RAGfly accesses the chunks it needs to retrieve and deliver context.
How is one customer's data isolated from another's?
Isolated workspaces by default: each workspace lives in a corpus isolated at the database level, with an Organization → Workspaces → Areas structure. A workspace is whatever you decide must never mix: your personal life and your business inside the same agent, or each client of your accounting practice. You draw the line. Nothing crosses it. Inside each workspace, you decide who sees what by area, role and agent. A consultancy or integrator can serve dozens of clients from a single platform without a single record crossing over.
Which formats are supported? Does it work with scanned PDFs (OCR)?
PDF, Word, Excel and plain text, including scanned documents thanks to first-rate OCR and table and layout understanding. The focus is text and complex documents, not audio or video.
Can I use it with regulated data (legal, healthcare, government)?
Yes. Desktop processes the original file and its extraction inside the customer's network. On sovereign plans, the customer owns the database where the derivatives persist. RAGfly still processes chunks on the way out to produce retrieval; a full in-situ runtime is the mode that keeps that processing inside the customer's infrastructure too.
Which AI models does it work with — Claude, GPT, Gemini?
RAGfly is model-agnostic. It delivers retrieval with citations and permissions; your AI —Claude, GPT, Gemini or another— uses RAGfly via MCP, REST or CLI.
How long does it take to go from a directory to your AI answering with citations?
You point RAGfly at a directory and, with automatic ingestion, vectorization and indexing, you can have your AI answering with citations in a short time, with no human help — and without building or maintaining a RAG pipeline at scale.
How does an AI agent consume RAGfly?
Through a remote MCP server (SSE or HTTP) or the RAGfly Desktop CLI, authenticating with a JWT or an API Key. The catalog of operations is published at `https://ragfly.ai/agents.json` and `https://ragfly.ai/llms-full.txt`.
Can I use my own vector database and my own model?
It depends on the plan. Growth enables parsing-BYOK, embedding-BYOK and agentic-BYOK; Scale adds a customer-owned Supabase vector project; Enterprise supports customer LLMs and compatible vector databases.
Not ready to sign up? Join the early-access list and we'll let you know.
Prefer to write to us? info@ragfly.ai