---
title: "Get Ai Suggestions"
method: POST
path: "/api/guidelines/ai-suggestions"
tags: ["Guidelines"]
---

# Get Ai Suggestions

`POST /api/guidelines/ai-suggestions`

## Request

- Base URL: `https://api.avete.ai` (the document declares no server; this is the origin it was published from)
- URL: `https://api.avete.ai/api/guidelines/ai-suggestions`
- Auth: HTTP bearer

Generate AI-powered suggestions for resolving violations.

Calls the ANZEN AI server's Gemini-powered /ai-suggestions endpoint to
generate LLM-driven corrective actions. There is intentionally no local
fallback: if the ANZEN server is unavailable or returns an error, an empty
result is returned so the UI surfaces a clear "suggestions unavailable"
state.

Non-empty results are cached in Redis (per report + violation + language)
with a TTL so that subsequent requests for the same report/violation are
served from cache without re-generating.

## Request body

- AISuggestionRequest
  - `guideline_ids` integer[], required
  - `language` string
  - `report_id` integer, nullable

## Response `200`

Successful Response

- unknown

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/avete/apis/fastcognito.md) - [All operations](https://skmtc.dev/avete/apis/fastcognito/llms.txt) - [OpenAPI document](https://skmtc.dev/avete/apis/fastcognito/revisions/2c3bbf575ddb?raw)