> 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/pipelines-with-venn.md).

# Pipelines with Venn

Draft telemetry pipeline rules from plain language, merge suggestions into existing pipelines, build pipelines in one step, and manage pipelines conversationally.

Venn turns intent into reviewable pipeline rules. Describe the goal; Venn samples your real logs, drafts the rules, and hands them to you in the pipeline builder for review. Nothing is saved until you press **Save**.

{% hint style="info" %}
For the full walkthrough with screenshots (the Review card, the pipeline builder preview, the DIFF tab, and per-rule-kind guidance), see [Venn and Pipeline build](/flow/venn-and-pipeline-build.md).
{% endhint %}

### What to ask for

* *"Reduce log volume from the checkout service"* (filter and suppression rules)
* *"Mask all email addresses in the auth logs"* (rewrite and redaction rules; the redaction is permanent, so check the before and after preview)
* *"Extract the status code and latency from my nginx logs"* (extract rules)
* *"Count errors per service every minute"* (aggregate rules)
* *"Forward only error logs to the SIEM destination"* (forward rules; a forward rule needs its own pipeline)
* *"Send audit logs to the security namespace"* (stream rules)

Venn picks the right rule kind for you, groups multiple rules into one pipeline where possible, and preserves any rules you wrote yourself when it refines a suggestion.

### Add rules to an existing pipeline

Ask Venn to extend a pipeline you already have:

* *"Add a rule to the payments pipeline that also drops health-check logs"*

Venn merges the new suggestions into the existing pipeline and deduplicates against what is already there, instead of creating a parallel pipeline.

### Manage pipelines conversationally

* *"List my pipelines"* (includes the savings each pipeline delivers)
* *"What does the drop-debug rule in the payments pipeline do?"* (Venn explains rules in plain language)
* *"Reorder the rules so the filter runs first"*, *"rename this pipeline"* (non-destructive structural edits: these act on the pipeline open in the builder, and you save the result yourself)
* *"Remove rule 3"* (a destructive edit; Venn asks for confirmation before removing the rule)
* *"Attach this pipeline to the prod dataflow"*, *"disable this pipeline"* (attaching, enabling, and disabling; replacing the pipelines on a dataflow asks for your confirmation first)
* *"Which dataflows have the most errors today?"* (health triage before you decide what to fix)

### The rule kinds Venn can draft

Filter, Rewrite (redaction), Code (JavaScript), Extract, Aggregate, Forward, Stream, and SIEM. One additional kind exists in the pipeline builder that Venn does not draft: **Tag** (also called Evaluate), which flags matched events and tracks them as a real-time metric; you author Tag rules by hand in the builder's **Add Rule** menu.