unbiased ai

RAGFlow, answering from your documents with Pareto

RAGFlow does the patient work: it parses your PDFs, splits them into chunks, finds the passages that matter and shows where each answer came from. The language model writes the answer. Add Pareto as an OpenAI-API-Compatible model and your chats and agents send that step to Pareto, which selects frontier and open models for each request and blends their outputs into one answer.

Choose Personal · $10/month See the setup →

Settings from RAGFlow’s docs50M tokens a week

An illustration of RAGFlow on Pareto. Not a screenshot.

Set it up as an OpenAI-API-Compatible model

RAGFlow talks to any service that speaks the OpenAI Chat Completions format through its OpenAI-API-Compatible provider. You create an instance with your Unbiased key, add pareto as a custom model, and make it the default LLM. These steps follow the current RAGFlow docs; older 0.x releases use different labels, covered in the FAQ below.

  1. Get an Unbiased API key

    Choose a plan, then create a key in the Unbiased platform. New accounts are reviewed and approved by a person before the key works.

  2. Create the provider instance

    Go to User settings > Model providers and pick OpenAI-API-Compatible under Available models. Enter an Instance name such as Unbiased, paste your key into API Key and set Base URL to https://api.unbiased.ai/v1. Keep the /v1: RAGFlow adds chat/completions itself.

  3. Add pareto as a custom model

    Click Add custom model, type pareto as the Model name and set Model type to Chat. Unbiased does not publish a model list, so type the name rather than waiting for it to appear. Leave Max context length at its default, or set 1000000. Turn on Tool calling only if your agents use tools, then click Confirm.

  4. Verify and save

    Click Verify all models, which checks each model against the instance’s key and Base URL, then save the instance. A failed check usually means a typo in the key or the Base URL. Unbiased does not reject unknown model names, so a check that passes does not prove you typed pareto correctly; check the spelling yourself.

  5. Make it the default LLM

    Open Set default models and choose pareto as the LLM. Keep an embedding model set as well; RAGFlow needs both. Then pick pareto in any chat assistant or Agent that should use it.

Settings follow RAGFlow’s own documentation: RAGFlow docs: Configure model providers.

Model providers > OpenAI-API-Compatible
Instance name:  Unbiased
API Key:        your-unbiased-api-key
Base URL:       https://api.unbiased.ai/v1

# Add custom model
Model name:          pareto
Model type:          Chat
Max context length:  default, or 1000000
Tool calling:        on only if agents use tools

# Set default models
LLM:  pareto

Pareto replaces the chat model only. Your embedding model, rerank model and the chunks already in your datasets stay as they are, so nothing needs to be parsed again. Format compatibility is not a promise that every RAGFlow feature works with Pareto.

The datasets stay. The writer changes.

Retrieval is RAGFlow’s job and it keeps doing it. Only the model that turns retrieved chunks into an answer is new.

One model string, several models behind it

Pareto is Unbiased’s blended AI model. On each request it selects frontier and open models and blends their outputs into one answer, so RAGFlow talks to one model string, pareto, and never has to pick.

“When the frontier improves, Pareto improves; when frontier prices fall, you should benefit.” From the pricing page.

How a Pareto answer gets made →
An illustration, not a wiring diagram. The blend includes frontier and open models, and its composition can change.

What an agent task costs on Pareto

A RAGFlow Agent can call the model several times before it answers, so the price of a finished task matters more than the price of a token. The model card publishes both for Pareto on two public agent benchmarks.

Terminal-Bench 4.0Tasks an agent completes in a terminal
$0.48/task
50.8% solved
DeepSWE v1.1Changes an agent makes in a real codebase
$0.24/task
69.9% solved

Pareto 26.10, preliminary October 1, 2026 runs that may change; denominators and the cost method still require confirmation. Cost is the reported mean USD per task at API rates, not an estimate for your workload. Model card and method.

Pay for RAGFlow by the month

One API key works across RAGFlow and any other tool you point at Pareto. Pick a monthly allowance, or pay per token on Team.

Personal

$10/month

  • 50M tokens a week
  • 20 requests per minute
  • For one person
Choose Personal

Personal Max

$100/month

  • 600M tokens a week
  • 60 requests per minute
  • For one person
Choose Personal Max

Team

Pay as you go

  • Prepaid credits at published rates
  • 300 requests per minute
  • No weekly token cap
Get started with Team

Subscriptions are for one person and are not unlimited. New API accounts are reviewed by a person before access turns on, and commercial or production use needs a separate written agreement. Rates and limits are on Pricing.

RAGFlow on Pareto, questions answered

Why isn’t pareto in RAGFlow’s model list?

RAGFlow fills the OpenAI-API-Compatible list by asking the provider which models it has, and Unbiased’s published docs say model listing is not offered. Use Add custom model and type pareto in lower case. RAGFlow’s docs ask that the name match the provider’s model ID exactly, and pareto is that ID.

Does the Base URL need /v1?

Yes. RAGFlow builds each chat request by adding chat/completions to the Base URL, so https://api.unbiased.ai/v1 becomes https://api.unbiased.ai/v1/chat/completions. RAGFlow does not add the /v1 for you, so it has to be in the field.

Can Pareto embed my documents too?

No. Pareto is a chat model, and RAGFlow needs a separate embedding model to index datasets and match questions to chunks. Keep the embedding model you use today; switching the LLM does not mean re-parsing anything.

My RAGFlow looks different. Where are these settings?

The steps above follow the current docs. In 0.x releases, open Model providers, find OpenAI-API-Compatible under Models to be added and click Add. Set Model type to chat, Model name to pareto, Base URL to https://api.unbiased.ai/v1 and your key in API Key. The Max tokens field there is the model’s context size, not an answer length. It defaults to 512, so set it to 1000000.

Can my team or customers use a RAGFlow app built on Pareto?

If RAGFlow serves your colleagues’ or customers’ questions in production, that is commercial use, which needs a separate written agreement; the platform teams page covers it. To switch a chat back, pick another model in its assistant settings.

Is Pareto made by the RAGFlow team?

No. Pareto is Unbiased’s model. RAGFlow is made by InfiniFlow, and Unbiased is not affiliated with or endorsed by them. You keep RAGFlow and point it at the Unbiased API.

What happens when I hit the weekly allowance?

Requests over your plan’s limits return an HTTP 429 error. Personal includes 50M tokens a week and 20 requests per minute. Personal Max raises that to 600M a week and 60 requests per minute for $100/month. Team uses prepaid pay-as-you-go credits with no weekly token cap.

Do I choose between models?

No. You send every request to one model string, pareto. Pareto selects frontier and open models for each request and blends their outputs into one answer.

Is my code kept?

Personal and Personal Max are not our zero data retention tier: prompts and responses may be kept up to 30 days, as section 5(c) of the Terms explains. For work that must not be kept, ask about the Zero Data Retention API on the Private AI page.

Pareto works in other tools too

The same key and the same model string go into each of these. See them all.

Product names belong to their owners. Except where a guide says otherwise, Unbiased is not affiliated with or endorsed by the makers of the tools on this page.

Ask your filing cabinet when the contracts renew

One provider instance, one custom model and the datasets you already parsed. Pareto starts at $10 a month for 50M tokens a week.

Choose Personal · $10/month See the setup →