> For the complete documentation index, see [llms.txt](https://docs.diaflow.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.diaflow.io/chat/tools.md).

# Tools

Tools are the operational building blocks and functional primitives that empower autonomous AI agents to interact with their environment, execute tasks, and process data. While language models provide reasoning and text understanding, Tools give agents the active ability to run code, query databases, browse the web, generate media, and modify workspace files.

### Core Tool Categories

1. **Data & Storage Primitives Tools** that allow agents to query, read, and mutate structured and unstructured data stored within the platform:
   1. Tables & Data (ddb): Inspect schemas, insert rows, query records, and manage structured tabular data dynamically. 
   2. Vector Search (vec): Index documents and perform semantic similarity searches across knowledge bases for retrieval-augmented workflows. 
   3. Drive & File Operations: Read, edit, transform, and write workspace files across multiple formats (PDF, DOCX, XLSX, PPTX, and Markdown). 

2. **Multi-Media & Asset Generation** Built-in execution tools that produce creative assets directly during conversational turns or workflow steps:

   1. Image Generation & Editing: Create high-resolution visual assets or perform targeted edits (such as background removal and style transfers). 
   2. Video Generation & Extension: Generate AI video clips from text or image prompts, and extend existing video sequences. 
   3. Speech & Audio Synthesis: Render text scripts into natural voiceovers and audio files automatically saved to workspace storage. 

   <figure><img src="https://3864624312-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FRdo1pkMnofwxf3lpWftw%2Fuploads%2Fuc1Zn2wgmy8JaIVvS9Lt%2Frounded-image%20(6).png?alt=media&amp;token=c6d262ba-7fbe-46d8-b2c2-33cacffdfa67" alt=""><figcaption></figcaption></figure>

3. **Execution & Computational Tools Capabilities** that give agents an active computing environment to run logic and analyze technical projects:
   1. Sandbox Bash & Script Execution: Run Python, shell scripts, and system commands in secure containerized sandboxes.
   2. Workflow & Node Execution: Trigger pre-built workflow graphs, execute specific nodes, and inspect flow schemas.&#x20;

4. **Search & External Egress Tools** that allow agents to gather real-time context and retrieve external content:

   1. Web Search & Fetch: Search public web sources, fetch page content, and render structured citations.&#x20;
   2. Academic & Media Search: Query DOI academic databases, retrieve research PDFs, and locate reference images online.

   <figure><img src="https://3864624312-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FRdo1pkMnofwxf3lpWftw%2Fuploads%2FMcSiTee3rStSvXqSbLLM%2Frounded-image%20(7).png?alt=media&amp;token=d3f8372a-d771-428f-88c9-942e65154901" alt=""><figcaption></figcaption></figure>

### How Tools Work in Diaflow&#x20;

* Autonomous AI agents automatically identify when a task requires tool execution based on the user's prompt or workflow requirements:
* Discovery & Schema Resolution: Agents inspect available tool schemas to understand parameters and expected inputs.&#x20;
* Permission & Governance: System rules ensure execution happens strictly within authorized sandbox boundaries and workspace scopes. Execution & Synthesis: The tool returns structured data, visual artifacts, or system outputs directly back to the agent for verification and final response composition.&#x20;

### Getting Started

* Default Availability: Standard tools (such as file handling, multi-media generation, and web search) are active by default inside Diaflow Chat and Studio.&#x20;
* Custom Tools: Developers can register custom function definitions and API tools to extend agent capabilities for specialized enterprise tasks.