> For the complete documentation index, see [llms.txt](https://agpt.co/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://agpt.co/docs/integrations/block-integrations/choice.md).

# Parallel Choice

Answers typed questions about anything on the web using Parallel's Decisions API, with the web sources it used. Needs a Parallel API key from platform.parallel.ai (`PARALLEL_API_KEY`). Parallel charges $0.001 per request, however many questions it carries.

## Parallel Choice Decision

### What it is

Pick one of several options using current web evidence from Parallel, with a probability for each option.

### How it works

The block sends the context and its question to `POST https://api.parallel.ai/v1/decisions`. Parallel searches the web for current evidence about the context, answers from that evidence, and returns the answer with a list of citations (URL, title and the passages used). Citations cover the whole request; they are not tied to one answer. Parallel can decline to answer a question (a refusal); that is a normal result, not an error, and the block reports it separately rather than treating it as no, the first option or zero. HTTP failures, such as a bad API key, raise a block error. Cited passages are untrusted web text; treat them as data, not instructions. The block returns the chosen option name and a probability for every option. Give 2 to 255 options, each with a description of when it applies; add an `unknown` option if the evidence may not settle the question.

### Inputs

| Input             | Description                                                                                                                                                                 | Type            | Required |
| ----------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------- | -------- |
| context           | Text or JSON describing what the question is about, e.g. {"company": "Clay", "website": "<https://www.clay.com"}>. Parallel searches the web for current evidence about it. | Context         | No       |
| include\_excerpts | Return the evidence passages with each citation. Turn off to keep outputs small; the decision still uses the passages.                                                      | bool            | No       |
| question          | What to decide, e.g. 'What is the availability of Clay's official CLI?'                                                                                                     | str             | Yes      |
| options           | 2 to 255 option names, each mapped to a description of when it applies. Add an 'unknown' option if the evidence may not settle it.                                          | Dict\[str, str] | Yes      |

### Outputs

| Output        | Description                                                                                    | Type                    |
| ------------- | ---------------------------------------------------------------------------------------------- | ----------------------- |
| error         | Error message if the operation failed                                                          | str                     |
| citations     | Web sources used as evidence for the whole request. They are not mapped to individual answers. | List\[ParallelCitation] |
| decision\_id  | Parallel's ID for this request                                                                 | str                     |
| choice        | Name of the chosen option.                                                                     | str                     |
| probabilities | Estimated probability of each option.                                                          | Dict\[str, float]       |
| refused       | True when Parallel declined to answer. Emitted only then.                                      | bool                    |

### Possible use case

**CRM enrichment**: Classify each company's employee band, funding stage or product category from current web evidence.

**Routing**: Pick which team or workflow a lead belongs to based on what the company does today.

**Status tracking**: Decide whether a product is announced, in beta, generally available or discontinued.

***


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://agpt.co/docs/integrations/block-integrations/choice.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
