> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tessact.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Multi-Agent Workflows

> Design repeatable, multi-team workflows that orchestrate first- and third-party AI agents, manual reviews, and automated transitions—at enterprise scale.

## Overview

Workflows are the backbone of collaboration in Tessact. A **workflow** defines *how work moves* from step to step across teams, which **AI agents** run (generative and analytical), and where **manual review layers** apply. You attach a workflow to one or more **projects** to run the same process consistently at scale.

<Frame caption="Workflow Builder canvas with steps, teams, and connections">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/workflow-canvas.png" alt="Workflow Builder canvas" />
</Frame>

**Key characteristics**

* **Composable steps**: Each step belongs to a team, has a **type**, and exposes **statuses** (with one default).
* **Semantic automation**: Status transitions can be triggered by **inputs**; completed statuses emit **outputs**. You wire outputs → inputs to automate orchestration across steps.
* **Multi-agent execution**: Invoke Tessact’s first-party agents (e.g., remix, compliance, captioning, thumbnail generation) or **bring your own** third-party AI via connectors and webhooks.
* **Safety by design**: Orchestration enforces **acyclic flows** (no infinite loops) and validates wiring before publish.

<Info>
  Attach the same workflow to many projects. Each project tracks its own run state, artifacts, and audit history without duplicating design work.
</Info>

***

## Core Concepts

### Steps, Types, and Statuses

* **Step** — A unit of work owned by a team (e.g., *Ingest*, *Edit*, *Compliance Review*, *Publish*).

* **Type** — Optimised configuration for that step’s work mode. Examples include:
  * **Watch Folders** — Monitor Connected Folders and react to file events.
  * **AI Agent** — Run one or more agents (e.g., *Generate YouTube thumbnails*, *Blog post from webinar*, *Compliance scan*).
  * **Video Creation** — Editor-centric step for timeline edits and remixes.
  * **Manual Review** — Gate with assignments, SLAs, and decision outcomes.
  * **Publishing** — Prepare deliverables and push to destinations/CMS.
  * **Utility / Transform** — Transcode, proxy, sprite sheet, metadata extraction.
  * **Webhook / External Action** — Call external systems and wait for callbacks.

* **Statuses** — Each step has a **default** status (e.g., *Not started*), plus additional statuses (e.g., *In progress*, *Needs edit*, *Approved*, *Rejected*). Statuses power both visibility and automation.

<Frame caption="Step configuration: team, type, default status, and additional statuses">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/step-config.png" alt="Step configuration" />
</Frame>

### Inputs & Outputs (Wiring)

* **Inputs** are *triggers* that cause a step’s status to change automatically (e.g., *AgentCompleted*, *FileMoved*, *ReviewApproved*, *TimeElapsed*, *WebhookReceived*, *MetadataUpdated*).
* **Outputs** are *signals* emitted when a status is reached (e.g., *Edit.Completed → outputs.Done*).

**Wiring rules**

* Connect **Output → Input** to propagate progress across steps.
* **Multiple Outputs → one Input** act as an **AND-gate**: the Input fires only when *all* connected Outputs are true.
* **One Output → multiple Inputs** (fan-out) is allowed.
* **No loops**: A step status cannot simultaneously use its own **IN** and **OUT** as triggers. The builder validates flows as a **DAG** (Directed Acyclic Graph).

<Warning>
  Recurring loops are disallowed. The builder prevents connections that would create cycles or self-triggering feedback.
</Warning>

<Frame caption="Status wiring with AND/OR patterns and validation">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/status-wiring.png" alt="Status wiring" />
</Frame>

***

## First-Party and Third-Party AI Agents

Tessact ships with first-party agents (remix, captioning, translation, compliance, summarisation, blog/title/thumbnail generation, etc.). You can also add third-party AI (LLMs, vision, audio) via **connectors** or **webhooks** and treat them as workflow steps.

<Tabs>
  <Tab title="First-party agent example">
    Configure a **Thumbnail Generator** agent:

    * **Inputs**: *Asset.Ready* (from Ingest), optional *MetadataUpdated:Brand*
    * **Parameters**: aspect ratios, brand palette, overlay template
    * **Outputs**: *Thumbnails.Created* (array of images), *Quality.OK/Fail*
  </Tab>

  <Tab title="Third-party agent example">
    Call an external captioning service:

    * **Webhook** step sends job spec (asset URL, language).
    * External system posts back to Tessact’s **Run Callback URL**.
    * On **Callback.Received** Input, move status to *Completed* and emit *Captions.Created*.
  </Tab>
</Tabs>

<Note>
  Agents can run in parallel within a step or as separate steps; use **conditions** to branch on their outputs (e.g., *Compliance.Fail → Manual Review*).
</Note>

***

## Designing Workflows

### Typical End-to-End Pattern

1. **Ingest (Watch Folders / Library)** → asset appears and analyses automatically.
2. **AI Agent Pack** → captioning, compliance, thumbnails, social remixes.
3. **Manual Review** → approval gate with assignments and SLAs.
4. **Publish** → deliver to CMS/social, archive to Connected Folder.

<Frame caption="Example: ingest → AI pack → review → publish">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/example-flow.png" alt="Example flow" />
</Frame>

### Advanced Patterns

* **Editorial factory**: *AI Search for Clips* → *Remix* → *QC* → *Publish*.
* **Knowledge expansion**: *Blog Generation from Video* → *SEO title & description* → *Manual copy edit* → *Publish*.
* **Compliance with regional overrides**: *Compliance EU* AND *Compliance US* → both must pass before *Publish*.

***

## Create a Workflow (Step-by-Step)

<Steps>
  <Step title="Open Workflow Builder">
    Go to **Workflows** in the sidebar → **New Workflow**.
  </Step>

  <Step title="Add steps">
    Drag **Step** onto the canvas. Choose **Type**, assign **Team**, and define **Statuses** (set a default).
  </Step>

  <Step title="Configure agents or actions">
    For **AI Agent** steps, select agents and parameters. For **Watch Folders**, choose a [Connected Folder](/connections/s3-setup) and event triggers.
  </Step>

  <Step title="Wire statuses">
    Connect **Output → Input** between steps. Use multiple Outputs to gate a single Input when all must be true.
  </Step>

  <Step title="Set manual review">
    Add a **Manual Review** step with assignment rules, due-by (SLA), and decision outcomes (*Approved*, *Needs edit*, *Rejected*).
  </Step>

  <Step title="Validate and publish">
    Click **Validate** to ensure the graph is acyclic and all required Inputs are satisfied. Then **Publish** the workflow.

    <Check>
      A green check indicates the workflow is ready to attach to projects.
    </Check>
  </Step>
</Steps>

<Frame caption="Validation panel with DAG check and unresolved inputs">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/validation-panel.png" alt="Validation panel" />
</Frame>

***

## Attach to a Project

* Create or open a **Project** → **Workflow** → select the published workflow.
* Provide any **step-level parameters** required by chosen types (e.g., which Connected Folder to watch, or agent presets).
* On save, the project enters the workflow’s **initial status** and begins execution when triggers fire.

<Info>
  A single workflow can be attached to many projects. Parameters are stored per project, not globally.
</Info>

***

## Manual Review Layers

* **Assignments** — Round-robin, owner-based, or group-based.
* **Checklists** — Define criteria (e.g., brand safety, loudness).
* **SLA timers** — Overdue states emit alerts and can auto-escalate.
* **Decisions** — *Approved*, *Needs edit*, *Rejected* emit Outputs you can wire forward.

<Frame caption="Manual Review step with assignees, checklist, and SLA">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/tessact/images/workflows/manual-review.png" alt="Manual Review UI" />
</Frame>

***

## Execution & Monitoring

* **Run History** — Timeline of step/status changes, agent jobs, and callbacks.
* **Retries & Backoff** — Transient failures auto-retry with exponential backoff; hard failures surface with actionable logs.
* **Idempotency** — External calls include a run-scoped idempotency key to avoid duplicate work.
* **Alerts & Webhooks** — Subscribe to events (e.g., *Step.Failed*, *Publish.Completed*).

```json Webhook (event subscription) theme={null}
{
  "event": "workflow.step.failed",
  "workflow_id": "wf_3bpv9",
  "project_id": "proj_9f2mx",
  "step": "Compliance Review",
  "status": "Failed",
  "attempt": 3,
  "run_id": "run_1a2b3c",
  "timestamp": "2025-08-25T10:45:12Z"
}
```

### Performance & Cost Management

* **Concurrency** — Limit concurrent agent executions per workflow or step.
* **GPU quotas** — Reserve or cap GPU-bound steps for predictable throughput.
* **Caching** — Reuse deterministic outputs (e.g., transcripts, embeddings) across runs.
* **Agent-compute minutes** — Usage accrues per agent runtime; see org-level analytics.

<Tip>
  Place heavy transforms behind success-gated compliance checks to avoid wasted compute on assets that will be rejected.
</Tip>

Troubleshooting

<AccordionGroup>
  <Accordion title="Validation fails: cyclic wiring">
    Remove or reroute connections creating loops. The DAG validator highlights offending edges.
  </Accordion>

  <Accordion title="Step waits forever for Input">
    Check upstream Outputs—if multiple are wired to the same Input, *all* must be true. Inspect the Run History for missing signals.
  </Accordion>

  <Accordion title="External webhook never returns">
    Verify the callback URL and idempotency key. Confirm firewall rules and TLS. Use the **Resend** action in Run History.
  </Accordion>

  <Accordion title="Agent exceeds quotas">
    Increase concurrency limits or schedule during off-peak hours. Review compute minutes and cache eligibility.
  </Accordion>
</AccordionGroup>

***

<Check>
  You’ve designed a robust, loop-safe workflow that orchestrates agents and people across projects. Attach it to a project to begin automated, traceable media operations at scale.
</Check>

***

## Next Steps

* [Create a Workflow](/workflows/create)
* [Attach to a Project](/workflows/attach)
* [Manual Review](/workflows/manual-review)