Rota Nacional product articles

Require privacy across your organization's API keys

Fri, 30 Oct 2026 13:49:50 GMT

# Require privacy across your organization's API keys

An organization can require the RN privacy barrier. Understand native API blocking, the 403 error and the protected chat path.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## When key consent is not enough

A team may need personal data to be processed before it reaches an external model. A connection's transmission consent must not remove that requirement. RN lets the organization record the requirement in its settings.

The choice covers that organization's keys and inference domains. It does not change other companies' settings. Requiring the barrier does not enable conversation storage: processing before transmission and content storage are separate choices.

## Enable the organization requirement

Select the organization in the dashboard. Open Settings, then Inference and privacy. Under Privacy barrier, select Edit, choose Require barrier and save. The change requires an organization owner or administrator session.

The management API field is policy.privacy_barrier_required. Updates send the version that was read, so an older edit cannot overwrite another change. Omitting the field preserves its previous value; it does not implicitly authorize disabling protection.

## What changes for integrations

When the requirement is enabled, /v1/providers/* returns 403 with code organization_privacy_barrier_required before processing the body or sending it to the provider. This also blocks files, batches and streaming on that path. RN does not attempt to transform a native format that its barrier cannot protect.

Use /v1/chat/completions with a rota/* model and the placeholder, redact or block policy. Monitor mode also returns 403 because it preserves the original content. If an application uses another organization's supplied pool, that pool's requirement also applies. Headers, prompts and key consent cannot remove it.

## Frequently asked questions

Does the barrier modify native API bodies?

No. When the organization requires the barrier, native requests are refused. Processing takes place on the protected chat path with the rota/* catalog.

Does this change other organizations' privacy settings?

The change stays with the selected organization. Another organization is affected only when it uses a pool supplied by the organization that imposed the requirement.

- [Current RN API contract](https://rota-nacional.ia.br/openapi.json) - [Privacy technology and limits](https://rota-nacional.ia.br/pt/tecnologia/)

[Open settings](https://rota-nacional.ia.br/en/app/settings/)

[Require privacy across your organization's API keys](https://rota-nacional.ia.br/en/articles/require-privacy-barrier/)

Before AI: replace personal data, redact it or block it?

Fri, 30 Oct 2026 13:49:50 GMT

# Before AI: replace personal data, redact it or block it?

Compare RN personal data policies and choose the right treatment for your workflow. Learn why monitoring is not anonymization.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Reduce the data sent to a model

A summary or classification request does not always need a person's name, phone number or identity document. On protected RN routes, the barrier examines content before the model call. It combines person-name detection with rules for personal data categories.

The purpose is to reduce exposure in the inference workflow. Detection has limits and is not a guarantee of irreversible anonymization. Validate results with representative fictional examples and retain your application's access controls.

## Choose the organization behavior

- placeholder replaces detected data with a marker, such as [PESSOA_1] or [CPF_1], to preserve some context. - redact removes the detected value. Use it when identification does not help the task. - block refuses the request with 422 when the barrier detects personal data. - monitor records detection but preserves content. It does not clean a request before transmission.

## Choose a path that supports the content

Protected chat uses the rota/* catalog. The /v1/privacy/clean endpoint processes text without calling a response model. For documents or images, check /v1/privacy/extract and capacity availability before sending content. A text barrier cannot directly inspect image pixels or audio.

The /v1/providers/* gateway preserves the original format and does not apply this processing. An organization can require the barrier and refuse that path. A request policy can only strengthen the organization's policy. Conversation storage, external transmission and personal data processing remain separate decisions.

## Frequently asked questions

Does monitor protect text sent to the provider?

No. Monitor preserves content. An organization that requires the barrier must use placeholder, redact or block on protected chat.

Can I use the barrier without generating an AI answer?

Yes. /v1/privacy/clean processes text without calling a response model. Check the API for access and payment conditions.

- [Rota Nacional technology](https://rota-nacional.ia.br/pt/tecnologia/) - [Privacy documentation](https://rota-nacional.ia.br/openapi.json)

[Try the barrier](https://rota-nacional.ia.br/en/app/try/)

[Before AI: replace personal data, redact it or block it?](https://rota-nacional.ia.br/en/articles/anonymize-data-before-ai/)

An AI gateway that preserves your application's model

Fri, 30 Oct 2026 13:49:50 GMT

# An AI gateway that preserves your application's model

Centralize AI connections in RN without converting provider request bodies. Understand native endpoints, authorized pools and migration limits.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Change access while preserving the contract

An application that already uses a provider may depend on parameters absent from a generic chat API. RN's native /v1/providers/{provider}/ gateway accepts the original API path. The model retains the format expected by the provider.

The application authenticates with an RN organization key. The external secret stays in the server connection. The gateway preserves the body, parameters, response and provider HTTP error. It does not substitute a rejected parameter or change the model to obtain an answer.

## Check three conditions before switching

- The connection must belong to the organization or an explicitly authorized supplier. - The application's path and contract must be supported by the gateway. Check files, streaming and asynchronous results as well as chat. - Native transmission requires consent and does not apply the content barrier. A privacy requirement from the organization or supplying pool refuses the request.

## Migrate one workflow at a time

Register the connection, check the financial contract and select the inference domain. Update the application's endpoint and credential. Validate a representative request, expected errors and resources read later by the workflow. A working chat does not prove batch or upload support.

Without X-RN-Connection-Id, selection takes place within the authorized pool. This header identifies a connection; it is not an access credential. Existing resources retain their original connection. The gateway does not automatically repeat a submitted call: check its state before trying again.

## Frequently asked questions

Does the application need the provider key?

The application uses its RN organization key. The provider secret stays encrypted in the server connection.

Does the native gateway anonymize requests?

No. It preserves the original body. Use protected chat for the barrier; when the barrier is mandatory, native requests return 403.

- [RN API index](https://rota-nacional.ia.br/api/) - [Gateway OpenAPI](https://rota-nacional.ia.br/openapi.json)

[Check the endpoints](https://rota-nacional.ia.br/en/docs/)

[An AI gateway that preserves your application's model](https://rota-nacional.ia.br/en/articles/native-gateway-keep-your-model/)

Manage multiple AI keys with organization isolation

Fri, 30 Oct 2026 13:49:50 GMT

# Manage multiple AI keys with organization isolation

Organize provider connections by company. Learn how pause, weights, consent and file ownership keep customer usage separate.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## A shared external account must not become shared access

When several applications use the same provider accounts, centralizing keys helps manage access. It does not authorize one company to read another's files, batches or responses. RN binds connections to the organization and records native resources in the space that created them.

The RN key authenticates the consumer. The external key belongs to the server connection. A connection identifier does not replace authentication or permit selection of another organization's key.

## Manage the pool without changing secrets in projects

A pool can contain multiple keys for the same provider. Administration supports pausing connections and setting distribution weights among eligible keys. Selection respects the authorized scope; it does not grant access to any organization's balance.

A paused or revoked owned connection does not automatically permit another pool. Inheritance requires a relationship, a contract and the relevant native consents. A required privacy barrier takes priority over native consent.

## Preserve asynchronous resource ownership

Files, batches and previous responses remain bound to their original connection. Later reads cannot randomly select another key. This matters when processing documents, tracking a batch or reading a task result.

Before revoking a connection, check pending work. When migrating an old resource, verify the organization, external account and original connection. Do not recreate paid work merely because its status lookup has not yet been integrated.

## Frequently asked questions

Can I register multiple keys for one provider?

Yes. A pool supports multiple connections, pause and distribution weights. Connections must be eligible for the organization and contract.

Does changing a weight move existing files and batches?

No. Existing resources keep their original connection. Weights guide selection for new eligible calls.

- [API scope and authentication](https://rota-nacional.ia.br/openapi.json) - [Rota Nacional documentation](https://rota-nacional.ia.br/pt/docs/)

[Manage the organization](https://rota-nacional.ia.br/en/app/settings/)

[Manage multiple AI keys with organization isolation](https://rota-nacional.ia.br/en/articles/provider-keys-organization-isolation/)

An AI API on your organization's own domain

Fri, 30 Oct 2026 13:49:50 GMT

# An AI API on your organization's own domain

Understand RN white label inference domains: branding, domain checks, HTTPS and binding to organization keys.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## The endpoint is part of an integration

A company can give its applications an inference endpoint under its own brand. RN binds the virtual domain to the organization. The endpoint keeps the API contract while its name and visual identity come from authorized settings.

This binding does not create shared customer access. One organization's domain refuses another organization's key. Branding does not transfer provider secrets to the browser or to public application code.

## From configuration to activation

The process includes registering the domain, checking its DNS destination, binding it to the correct organization and enabling HTTPS. Configure the specified records and wait for validation. Do not replace an application's endpoint until the domain is active and its certificate works.

The organization manages its brand name, logo and legal information. Email delivery has separate settings. Enabling a brand does not prove that its sender is authorized to deliver email; check that state separately.

## Keep policies when changing the address

A call through a custom domain still follows organization permissions, allowed networks, quotas and the financial contract. Native access still requires consent. When the organization requires the barrier, changing the hostname does not unlock /v1/providers/*.

Before switching, validate the correct key, the application's API path and access to existing resources. Keep a rollback path to the previous endpoint. Do not publish customer names, domains or credentials in marketing material without authorization.

## Frequently asked questions

Does a custom domain accept another company's key?

No. The domain is bound to the organization and authentication must match it.

Does an active brand mean email is also active?

No. Branding and sender configuration have separate states. Delivery depends on email configuration and sender authorization.

- [RN public contract](https://rota-nacional.ia.br/openapi.json) - [Rota Nacional integrations](https://rota-nacional.ia.br/pt/docs/agentes/)

[Open your organization](https://rota-nacional.ia.br/en/app/settings/)

[An AI API on your organization's own domain](https://rota-nacional.ia.br/en/articles/custom-domain-white-label-ai-api/)

Track inference costs, limits and receipts

Fri, 30 Oct 2026 13:49:50 GMT

# Track inference costs, limits and receipts

Distinguish reservations, measured usage and charges. Read RN receipts, interpret pending states and avoid repeating paid calls.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Tokens do not describe the whole bill

A request may reserve funds before the answer and finish with lower consumption. The reservation protects the authorized ceiling. Settlement uses the contract and measurable usage. Reserved and charged amounts therefore appear as separate information.

Provider cost and RN service cost are also different. An owned key and a supplied key may follow different contracts. Check the tariff that applies to your case. A missing value does not mean zero cost.

## Recover state before repeating work

When the gateway creates a reservation, X-RN-Inference-Receipt identifies its receipt. The /v1/inference/receipts route reads the payer's authorized records. Receipts are also available in inference management.

An interrupted connection does not prove that the provider did not execute the request. Unknown usage can remain pending reconciliation. Check the receipt and asynchronous resource state before submitting the same work again. RN does not automatically repeat a submitted native request.

## Set limits at more than one level

An output limit bounds the requested answer size. Organization quotas control requests and tokens over a period. Balance and contract control financial authorization. These controls complement one another; none replaces the others.

Files and reads are not equivalent to a new inference. Batches need a contract that accounts for later processing; a token tariff at creation does not measure all batch work. When investigating an error, retain request and receipt identifiers without putting prompts or keys in public logs.

## Frequently asked questions

Are reserved and charged amounts the same?

Not necessarily. A reservation covers the authorized ceiling. Settlement uses the contract and measured usage; the difference can be returned.

Does a connection failure mean I should repeat the call?

Not by itself. The provider may have received the request. Check its receipt and work state before trying again.

- [API receipts and limits](https://rota-nacional.ia.br/openapi.json) - [Pricing and access](https://rota-nacional.ia.br/api/pricing)

[View my usage](https://rota-nacional.ia.br/en/app/usage/)

[Track inference costs, limits and receipts](https://rota-nacional.ia.br/en/articles/inference-receipts-limits-costs/)

Try AI with verified email and a human check

Fri, 30 Oct 2026 13:49:50 GMT

# Try AI with verified email and a human check

Learn how to start with RN, verify your account and use the available trial. Understand quotas, the first-use human check and continuation with credit.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Start with a small task

Open RN chat and choose a task you can assess: summarize fictional text, review a sentence or organize a list. Sign in with a verified email account or Google. Complete the acceptance and human check before the free allowance is granted.

Validation takes place on the server. Opening a window or knowing the endpoint does not grant the benefit. The system checks session identity, verified email and the human verification response.

## Free usage has limits

Check current amounts at /api/pricing and in the dashboard. The new-account trial and any daily quota follow separate rules. Available balance accounts for reservations from requests still running. An enabled offer does not mean unlimited use or permanent availability of every provider.

The initial allowance is per person. Switching organizations does not create another. A daily quota depends on active configuration and the platform's shared ceiling. An account with purchased credit does not need a new grant to use its paid balance.

## Continue when the allowance ends

If a request exceeds available balance or quota, RN explains the condition and offers credit purchase in the workflow. The payment dialog opens over the task. After confirmation, the user continues without reconstructing the conversation.

Free website access does not make the entire API free. Paid capabilities and agents follow the conditions of their endpoint. Start with an available feature and check the price before authorizing a paid call.

## Frequently asked questions

Can I get another trial by creating an organization?

No. The initial allowance is per person. Switching organizations does not renew it.

Do I need a verified email for free usage?

Yes. The account needs verified email or Google sign-in, plus acceptance and the human check required for the grant.

- [Current trial conditions](https://rota-nacional.ia.br/api/pricing) - [API account and allowance](https://rota-nacional.ia.br/openapi.json)

[Try a conversation](https://rota-nacional.ia.br/app/bots/)

[Try AI with verified email and a human check](https://rota-nacional.ia.br/en/articles/free-ai-trial-verified-email/)

AI credit by Pix: continue after your trial

Fri, 30 Oct 2026 13:49:50 GMT

# AI credit by Pix: continue after your trial

Understand RN credit purchase, organization balance and usage charges. Payment opens over the task and preserves your work context.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Buy for the account that will use the credit

RN offers continuation with prepaid credit. When a task requires balance, checkout shows the amount and conditions before confirmation. Buying one-time credit does not implicitly start a subscription.

Check the selected organization before paying. Personal account balance and organization balance are separate usage wallets. An organization's keys follow its context, permissions and consumption conditions.

## Payment follows the task

Balance and purchase open in a dialog over your work. You do not have to abandon the conversation to find a sales page. After payment confirmation, the product updates the balance and resumes the usage path.

Use the state reported by RN. A receipt image or an open payment screen does not replace confirmation by the financial workflow. If Pix is pending, track that state before creating another charge.

## Check the selected feature's price

Models and capabilities use different billing units. Chat may charge input and output tokens; other capabilities follow their published unit. A reservation can cover an answer's ceiling and be adjusted later using measured consumption.

Check prices and receipts before automating large volumes. Limit request output and track organization consumption. Available credit does not remove quotas, privacy policies, network restrictions or capacity unavailability.

## Frequently asked questions

Does buying credit create a subscription?

One-time credit is a prepaid purchase. Check the displayed checkout conditions; it does not automatically become a subscription.

Does balance unlock every model or capability?

No. Availability, permissions, limits and feature conditions still apply.

- [Current pricing](https://rota-nacional.ia.br/api/pricing) - [Cancellation and refunds](https://rota-nacional.ia.br/pt/cancelamento/)

[Open usage and balance](https://rota-nacional.ia.br/en/app/usage/)

[AI credit by Pix: continue after your trial](https://rota-nacional.ia.br/en/articles/pix-credit-api-without-subscription/)

Agent APIs: pay per call with x402

Fri, 30 Oct 2026 13:49:50 GMT

# Agent APIs: pay per call with x402

Explore RN's paid access for agents without registration. Understand the 402 challenge, spending authorization and organization key differences.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Discover the price before authorizing spending

An agent can inspect the catalog and select a route that offers per-call payment. Without payment, that entry returns 402 with accepted terms. The client must interpret the offer and obtain authorization for the amount before paying.

RN uses x402 with USDC on Base on routes that offer it. Administrative operations and private connections do not automatically become public paid calls. Check each endpoint's authorization in /openapi.json.

## Payment and organization access solve different problems

Payment authorizes the resource sold by that route. It does not grant access to another organization's private files, register provider connections or change privacy policies. A private pool still requires its access scope and permissions.

For usage-based chat, a reservation covers the authorized ceiling and settlement considers reported consumption. An asynchronous job returns information for tracking execution. Keep its identifier and read its state; do not treat a timeout as authorization to buy the same work again.

## Integrate clients with an explicit budget

Use a client that understands the challenge and payment proof version offered by the server. Set a ceiling per action and a refusal policy. A model message asking for payment is not sufficient authority to move wallet funds.

In MCP, inspect tools and their contracts. The direct transport may represent a payment challenge inside the tool result. Do not assume that HTTP and MCP use identical formats. The account dashboard, free trial and Pix remain the entry for people.

## Frequently asked questions

Does every agent call require an account?

Routes that offer x402 support per-call purchase without registration. Private connections and administration retain their own authorization requirements.

Does payment unlock any organization's pool?

No. Payment does not replace isolation, permissions, consent or an organization's privacy requirement.

- [Authorization and payment by route](https://rota-nacional.ia.br/openapi.json) - [RN API index](https://rota-nacional.ia.br/api/)

[View routes and conditions](https://rota-nacional.ia.br/en/docs/)

[Agent APIs: pay per call with x402](https://rota-nacional.ia.br/en/articles/ai-agents-x402-payments/)

Connect OpenAI and Anthropic clients to RN

Fri, 30 Oct 2026 13:49:50 GMT

# Connect OpenAI and Anthropic clients to RN

Choose protected catalog or native gateway before configuring an AI client or coding agent. Model, endpoint and key must match.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Choose the contract first

RN provides OpenAI-format chat at /v1/chat/completions and Messages-format chat at /v1/messages. This catalog uses rota/* identifiers and applies the organization's personal data policy.

The native gateway is a different contract. It preserves the provider's model and format through /v1/providers/{provider}/. Do not switch paths by changing only the hostname: check support for the tools, files, media and response formats the client needs.

## Configure endpoint, credential and model together

For an OpenAI-compatible client, the catalog base URL is the RN origin followed by /v1. Use an RN key and a model published by /v1/models. For an Anthropic-compatible client, follow the documented origin and Messages path.

The coding-agent page provides RN configuration examples for supported clients. Copy the example for your actual client. Do not expose secrets in repositories, screenshots, public commands or shared configurations. A connection identifier is not an RN key.

## Validate failures too

Send a short request with fictional content. Then check the streaming, tools and limits your workflow actually uses. A 401 indicates a credential problem; a 403 may come from organization policy; a 402 requires balance or valid financial terms. The error body provides the specific code.

When the organization requires the barrier, the client must use a protected path. A coding agent does not gain permission to send raw content by selecting another provider. The application retains control of which files and information enter its context.

## Frequently asked questions

Can I use any model name in the protected catalog?

Use identifiers published by /v1/models and enabled for the organization. Provider-native names belong to the native gateway contract.

Does chat compatibility cover the provider's entire API?

No. Check the paths and features your application uses. Files, media, tools and asynchronous operations have their own contracts.

- [Agent configuration](https://rota-nacional.ia.br/pt/docs/agentes/) - [Public model catalog](https://rota-nacional.ia.br/v1/models)

[Configure my client](https://rota-nacional.ia.br/en/docs/agents/)

[Connect OpenAI and Anthropic clients to RN](https://rota-nacional.ia.br/en/articles/connect-openai-anthropic-clients/)

n8n with AI and privacy: a customer service workflow

Fri, 30 Oct 2026 13:49:50 GMT

# n8n with AI and privacy: a customer service workflow

Use RN's n8n example to receive a message and return an answer. Check authentication, limits and responsibilities before connecting your channel.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## The example performs a bounded task

RN provides an importable n8n customer service workflow. It receives a message field from another workflow, validates input and calls RN chat. The result contains reply and model. The example uses fictional data so the first run needs no customer information.

The workflow does not send WhatsApp messages. The calling application owns the channel. It must authenticate input, control duplicates and decide when to hand the conversation to a person.

## Separate authentication from content

Import the integration JSON and configure Bearer Auth in the HTTP Request node. Keep the key in n8n credentials, separate from the JSON body. The example does not need an external provider key: it uses the RN key.

Check the rota/* model and the key's organization. The call follows RN's configured policy. If the team requires the barrier, continue using the protected contract when adapting the workflow.

## Connect to a real process with controls

The published version validates messages from 1 to 8,000 characters. It does not save execution data or automatically retry a failed call. Preserve a retention policy appropriate to your process when adapting the workflow.

Test errors, insufficient balance and human handoff before connecting a real channel. Do not treat a generated answer as authority to issue charges, change records or promise commercial terms. Those actions require application rules and permissions.

## Frequently asked questions

Does the workflow send the answer to WhatsApp?

No. It returns reply and model. The calling application integrates the channel and controls authentication, duplicates and human handoff.

Does the key belong in the request JSON?

No. Configure Bearer Auth in n8n and keep credentials separate from content.

- [n8n workflow JSON](https://rota-nacional.ia.br/integracoes/rota-atendimento-n8n.json) - [Integration guide](https://rota-nacional.ia.br/pt/solucoes/agentes-de-whatsapp/#n8n)

[Open the n8n workflow](https://rota-nacional.ia.br/en/solutions/whatsapp-agents/#n8n)

[n8n with AI and privacy: a customer service workflow](https://rota-nacional.ia.br/en/articles/n8n-customer-service-privacy/)

Native OCR or privacy extraction: which path fits?

Fri, 30 Oct 2026 13:49:50 GMT

# Native OCR or privacy extraction: which path fits?

Compare native Mistral OCR with RN text extraction and personal data processing. Input type and policy change the integration.

Rota Nacional · 2026-10-30T13:49:50Z · 2026-10-30T13:49:50Z

## Reading a document is not the same as protecting its text

OCR converts visual content into text and structure. The privacy barrier processes personal data in content it can inspect. An image sent directly to a model is not protected merely because another part of the conversation passed through text rules.

Choose the path before integrating. If the organization requires processing before transmission, check protected extraction and available capabilities. If you need a provider-native contract, assess authorization to send the original document.

## When to use native OCR

The /v1/providers/mistral/v1/ocr path preserves the body accepted by Mistral, including the model and supported provider options. Input can reference a document or image in the native format, or an owned file registered in the gateway.

This path requires an authorized connection and native consent. It sends the original document without the privacy barrier. An organization that requires the barrier receives 403 before transmission. The gateway preserves provider errors rather than modifying the request to make it pass.

## When to use protected extraction

The /v1/privacy/extract endpoint is the entry for obtaining processed text before using protected chat. Check supported formats, maximum size and capacity availability in the API. The existence of a route does not mean a worker is available at that moment.

Start with a representative fictional document. Check reading order, tables and detected categories. OCR can misread characters and structure; personal data detection has limits too. A successful response does not establish perfect extraction or automatic legal compliance.

## Frequently asked questions

Does native OCR apply the privacy barrier?

No. It preserves the document and provider contract. When the organization requires the barrier, this path is refused.

Can I put an image in chat for the barrier to read?

The text barrier cannot inspect pixels. Check protected extraction, supported formats and capacity availability before sending the content to chat.

- [RN routes and limits](https://rota-nacional.ia.br/openapi.json) - [Official Mistral OCR contract](https://docs.mistral.ai/api/endpoint/ocr)

[Check extraction and OCR](https://rota-nacional.ia.br/en/docs/)

[Native OCR or privacy extraction: which path fits?](https://rota-nacional.ia.br/en/articles/native-ocr-private-extraction/)