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

# Action Reference

> Every action available on the Workflow canvas, what it does, and where it applies.

Actions are the building blocks of a [Workflow](/features/workflows). Each action runs an operation on an input and produces an output — usually a Corvic Table or a Space. This page is a quick reference for every action, what it applies to, and where to learn more.

## Actions by input type

### Data sources

Actions that run on a raw data source.

| Action                           | Applies to   | Output                           | Learn more                                         |
| -------------------------------- | ------------ | -------------------------------- | -------------------------------------------------- |
| Multi-modal Knowledge Extraction | Unstructured | Tables (text, structure, images) | [Guide](/features/multimodal-knowledge-extraction) |
| Document Digitization (OCR)      | Unstructured | Markdown / text                  | [Guide](/features/document-digitization)           |
| Extract Tables                   | Structured   | One table per schema             | [Guide](/features/extract-tables)                  |
| Extract Images                   | Unstructured | Image table                      | [Guide](/features/extract-images)                  |
| Apply Template                   | Data source  | A full pre-built pipeline        | [Workflows](/features/workflows)                   |
| Web Search ingestion             | Web source   | Structured table                 | [Data Sources](/features/data-sources)             |

### Corvic Tables

Actions that transform or combine Corvic Tables.

| Action             | Applies to           | Output             | Learn more                       |
| ------------------ | -------------------- | ------------------ | -------------------------------- |
| Augment (AI / LLM) | Corvic Table         | Augmented table    | [Guide](/features/llm-augment)   |
| Web Augment        | Corvic Table         | Augmented table    | [Guide](/features/web-augment)   |
| Code Augment       | Corvic Table         | Augmented table    | [Guide](/features/code-augment)  |
| Join               | Two or more tables   | Joined table       | [Guide](/features/join)          |
| Graph Builder      | Two or more tables   | Graph table        | [Guide](/features/graph-builder) |
| Python             | Corvic Table / Space | Transformed output | [Guide](/features/python-sql)    |
| SQL                | Corvic Table / Space | Transformed output | [Guide](/features/python-sql)    |
| Embed              | Corvic Table         | Space              | [Guide](/features/embed)         |

### Spaces

Actions that run on embedding Spaces.

| Action       | Applies to | Output             | Learn more                                |
| ------------ | ---------- | ------------------ | ----------------------------------------- |
| Create Agent | Space      | Agent              | [Guide](/features/workflows#create-agent) |
| Python       | Space      | Transformed output | [Guide](/features/python-sql)             |
| SQL          | Space      | Transformed output | [Guide](/features/python-sql)             |

## Embedding types

The **Embed** action can produce several types of Space, depending on your data:

| Type           | Use for                                                |
| -------------- | ------------------------------------------------------ |
| **Semantic**   | Text — semantic search and RAG                         |
| **Tabular**    | Structured/tabular encoding                            |
| **Image**      | Image embeddings                                       |
| **Relational** | Graph/relational structure (often after Graph Builder) |

See [Embed](/features/embed) and [Graph Builder](/features/graph-builder) for details.

## UI label ↔ capability

For reference, the user-facing action labels map to these underlying capabilities:

| UI label                         | Capability                                              |
| -------------------------------- | ------------------------------------------------------- |
| Multi-modal Knowledge Extraction | `parse`                                                 |
| Document Digitization            | `convert_to_markdown`                                   |
| Extract Tables                   | `stack_parquet` / `demux_database_tables`               |
| Extract Images                   | `extract_images`                                        |
| Augment (AI)                     | `llm_augment`                                           |
| Web Augment                      | `web_augment`                                           |
| Code Augment                     | `code_augment`                                          |
| Join                             | `join`                                                  |
| Build Graph                      | `build_graph` / `graph_embed`                           |
| Embed                            | `concat_and_embed` / `embed_and_concat` / `embed_image` |
| Python                           | `execute_python`                                        |
| SQL                              | `execute_sql`                                           |
| Create Agent                     | `create_agent`                                          |

<Note>
  Available actions are context-aware: the **+** menu on a node only shows actions valid for that node's type. Some capabilities are used internally by templates and may not appear directly in the menu.
</Note>

## Related Documentation

<CardGroup cols={2}>
  <Card title="Workflows" icon="window" href="/features/workflows">
    Build pipelines from these actions.
  </Card>

  <Card title="Builder Chat" icon="wand-magic-sparkles" href="/features/builder-chat">
    Let the assistant assemble actions for you.
  </Card>

  <Card title="Corvic Tables" icon="table-cells" href="/concepts/feature-views">
    The tables most actions produce.
  </Card>

  <Card title="Spaces" icon="layer-group" href="/concepts/spaces">
    The embeddings the Embed action produces.
  </Card>
</CardGroup>