> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.ada.cx/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ada.cx/_mcp/server.

# Articles

**Articles** represent a unit of content within the knowledge source, containing information that an AI Agent can use to resolve an end user inquiry.

## Availability rules

The `availability_rules` field on an article controls which articles the AI Agent can access during a conversation. When a rule is set, the AI Agent can only retrieve the article on conversations where the variables' values match the rule. See [Availability rules](/reference/introduction/availability-rules) for the rule schema, operators, validation errors, and examples.

### Field semantics on upsert

| Value         | Effect                                                                                                                  |
| ------------- | ----------------------------------------------------------------------------------------------------------------------- |
| Rule object   | Attach (or replace) the rule on the article.                                                                            |
| `null`        | Detach any existing rule; the article becomes available to the AI Agent for all conversations.                          |
| Field omitted | Leave the existing rule unchanged. On a new article, apply the availability rules of its source, if the source has any. |

### Field semantics on read

`availability_rules` is a rule object when an availability rule is set, or `null` when no rule is attached.

## Performance recommendations

For best performance on `POST /v2/knowledge/bulk/articles/`:

* Keep rules concise - use only the conditions your targeting needs.
* You may see degraded performance on requests with more than 5,000 total conditions across all articles.
* If you're setting rules across many articles, split the work into multiple requests rather than one large batch.