canonical: https://jentic.com/apis/messagebird.com/messagebird-faq-api

# MessageBird FAQ

Jentic publishes the only available OpenAPI specification for MessageBird FAQ, keeping it validated and agent-ready. The MessageBird FAQ API hosts curated question-and-answer datasets and predicts the best matching answer for an end-user question against a chosen dataset. It exposes dataset listing and retrieval, an answer-prediction endpoint that scores candidate answers from the dataset, and a supported-languages lookup, making it a focused building block for FAQ-driven chat and IVR experiences inside the wider MessageBird platform.

## For AI agents

Predict the best FAQ answer to a user question against a curated MessageBird dataset across multiple languages.

## Scope

Does not handle SMS or voice messaging, free-form text generation, or dataset authoring - use for FAQ dataset retrieval and answer prediction only.

## Capabilities

- List FAQ datasets configured for the MessageBird account
- Retrieve a specific FAQ dataset by ID for inspection or content syncing
- Predict the best matching answer for a free-text user question against a chosen dataset
- Look up which languages are supported for FAQ prediction before routing a query
- Power chatbot and IVR flows with grounded FAQ responses rather than free-form generation

## Use cases

### FAQ-Grounded Chatbot

Customer-support chatbots can ground their responses in a curated FAQ dataset by calling POST /api/v1/datasets/{datasetId}/predict-answer with the user's question. The endpoint returns the best matching answer from the dataset along with confidence information, so the bot can reply with a vetted answer or escalate when the score is too low.

Example prompt: Call POST /api/v1/datasets/{datasetId}/predict-answer with the customer's question and serve the predicted answer if its confidence exceeds the configured threshold.

### IVR Self-Service

Voice teams running IVR flows on top of MessageBird can use the FAQ API to handle natural-language questions captured by speech-to-text. A successful prediction lets the IVR speak the FAQ answer back; a low-confidence response triggers transfer to a human agent. The supported-languages endpoint helps route per-locale FAQs.

Example prompt: Call GET /api/v1/supported-languages, then POST /api/v1/datasets/{datasetId}/predict-answer for the caller's locale to drive the IVR self-service step.

### Knowledge Base Sync

Internal knowledge platforms can keep their content aligned with the MessageBird FAQ datasets by listing datasets via GET /api/v1/datasets and pulling individual datasets via GET /api/v1/datasets/{datasetId}. This supports a publishing workflow where editors maintain answers in a CMS and downstream chat or voice channels read from the synced dataset.

Example prompt: Call GET /api/v1/datasets to list datasets, then GET /api/v1/datasets/{datasetId} to pull the active dataset into the internal knowledge base.

### Agent-Driven Customer Support

An AI agent fielding inbound customer messages can call the MessageBird FAQ API via Jentic to ground its replies in vetted answers. The agent searches Jentic for the predict-answer operation, loads the schema, and posts the user's question without writing custom MessageBird client code or handling raw access keys.

Example prompt: Search Jentic for 'predict the best FAQ answer with MessageBird', load POST /api/v1/datasets/{datasetId}/predict-answer, and execute it with the inbound user question.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| GET | /api/v1/datasets | List FAQ datasets |
| GET | /api/v1/datasets/{datasetId} | Retrieve a specific FAQ dataset |
| POST | /api/v1/datasets/{datasetId}/predict-answer | Predict the best matching answer |
| GET | /api/v1/supported-languages | List supported languages |

## Key resources

- **Datasets** — FAQ dataset catalogue and individual dataset retrieval
- **Predict Answer** — Score and return the best matching answer for a free-text question
- **Supported Languages** — Language coverage for FAQ prediction

## Why Jentic

- **Setup:** Wiring the MessageBird FAQ API by hand means attaching your AccessKey to each request against the faq.messagebird host, routing dataset lookups and answer prediction, and handling retries yourself. Through Jentic you install once, import the FAQ API from the API Directory, store the AccessKey once, and your agent calls it.
- **Permission scoping:** The FAQ API puts the dataset id in the URL path (/api/v1/datasets/{datasetId}/predict-answer), so a rule can pin your agent to one dataset. You choose the operations it may call, so a rule can grant dataset reads and answer prediction while leaving out anything else.
- **Credential handling:** Your MessageBird AccessKey is stored once, encrypted, by your own Jentic One instance and attached at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'predict an FAQ answer' or 'list FAQ datasets', and Jentic returns the matching operation such as POST /api/v1/datasets/{datasetId}/predict-answer with its dataset id and question fields so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **MessageBird Intent API** — Detect user intent before routing to the right FAQ dataset
- **MessageBird Language Detection API** — Detect the user's language to choose the correct localised FAQ
- **OpenAI API** — General-purpose LLM for free-form answer generation

## FAQ

### Why is there no official OpenAPI spec for MessageBird FAQ?

MessageBird does not publish an OpenAPI specification for the FAQ product. Jentic generates and maintains this spec so that AI agents and developers can call MessageBird FAQ via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the MessageBird FAQ API use?

The FAQ API uses a MessageBird AccessKey passed in the Authorization header. Through Jentic, that AccessKey lives in the credential vault and the agent calls the API with a scoped Jentic token, so the raw key never enters the agent context.

### Can I get a confidence score with the predicted answer?

Yes. POST /api/v1/datasets/{datasetId}/predict-answer returns the best-matching answer from the dataset along with score information so callers can apply a confidence threshold and fall back to human handoff when the match is weak.

### What are the rate limits for the MessageBird FAQ API?

The OpenAPI spec does not declare explicit rate limits. MessageBird applies account-level throttling on FAQ endpoints; high-volume integrations should batch retrieval calls and back off on 429 responses rather than calling predict-answer in a tight loop.

### How do I predict an FAQ answer through Jentic?

Install the SDK with pip install jentic, search for 'predict the best FAQ answer with MessageBird', load the POST /api/v1/datasets/{datasetId}/predict-answer operation, and execute it with the dataset ID and the user question. Jentic returns the matched answer and score.

### Which languages does the MessageBird FAQ API support?

Call GET /api/v1/supported-languages to retrieve the current list. The dataset configuration determines which of those languages a specific FAQ can be queried in, so route queries by checking the dataset's locale before predicting.

### Can I limit what my agent is allowed to do with the MessageBird FAQ API?

Yes. Because you run Jentic One yourself, your own rules decide which MessageBird FAQ operations the agent may call, so you can grant only dataset reads (GET /api/v1/datasets and GET /api/v1/datasets/{datasetId}) and answer prediction (POST /api/v1/datasets/{datasetId}/predict-answer) while withholding everything else. Since the dataset id sits in the URL path, a rule can pin the agent to a single FAQ dataset rather than every dataset on the account. Your stored MessageBird AccessKey is attached only at execution time, so the agent never sees the raw credential.