> For the complete documentation index, see [llms.txt](https://docs.apica.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.apica.io/venn-ai-assistant/getting-started-with-venn.md).
# Getting Started with Venn
Where Venn lives in Apica Ascent, how to start a conversation, and what to ask first.
Venn is available on almost every page of Ascent. You never need to go looking for it.
### Three ways to open Venn
* **Floating bubble**: click the bubble in the corner of any page (tooltip **"Ask Venn"**). It opens the Venn drawer.
* **Drawer**: a panel slides in from the right, titled **Venn**. You can pin it, widen it, start a **New chat**, or close it. The drawer stays with you as you navigate, and it is aware of the page you are on.
* **Full page**: choose **Venn** from the top navbar, or simply log in: the full-page chat at `/venn` is the default landing page. It shows **"Ask Venn anything"**, starter prompts, and your conversation history.
The "Ask Venn" bubble
The Venn drawer's empty state
Type your question and press **Enter** to send (**Shift+Enter** for a new line).
### Your first prompts
Good first questions to get a feel for Venn:
* *"What namespaces am I ingesting data from?"*
* *"Show me errors from the last hour"*
* *"Which dataflows have the most errors in the last 24 hours?"*
* *"How do I set up a Kafka source?"* (Venn answers how-to questions from the Apica documentation, inline in the chat)
* *"Help me cut log volume from my noisiest app using telemetry pipelines"*
### Conversations
* Every chat is saved under **Conversations** on the full-page view, so you can return to and continue past conversations.
* Switching to another chat does not stop Venn: an in-flight answer keeps generating in the background, and results such as pipeline Review cards stay usable when you come back.
* After many answers Venn offers follow-up suggestions as clickable buttons. Clicking one sends it as your next message; you can always type your own instead.
### First-time setup: connect an AI provider (admins)
Venn needs an AI provider before it can answer. An administrator sets this up once, in **Settings > Admin Settings > Configuration**, under the **AI & Model** section:
1. Pick the **AI Provider**: **Anthropic**, **OpenAI**, **Azure**, or **LocalAI** (a self-hosted OpenAI-compatible endpoint).
2. Enter the provider's **API key** and **model** (plus endpoint or base URL where applicable). Keys are stored as secrets and shown masked once saved.
3. Click **Update Configuration**. Changes apply immediately: the next message Venn handles uses the new provider. No restart is needed.
The same section also holds Venn's agent limits and switches (tool-call budgets, prompt caching, delete-tool enablement). See Administer Venn for the full reference.
### What Venn will and will not do
Venn is an observability assistant, and it stays one:
* It answers questions about your logs, metrics, checks, dataflows, and about the Ascent product itself.
* It declines requests outside that scope (general knowledge, politics, unrelated coding tasks, legal or medical advice) and steers back to your data.
* It drafts things for you to review; it does not save or delete anything without your explicit approval. See Safety and control.
