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

# Overview

Tools extend agent capabilities beyond text generation. NeuronAI Studio supports database-managed tool types plus built-in toolkits, scanned PHP classes, MCP-exposed tools, and workflow Tool Mode specialists (`NodeAsTool`).

## Tool sources

```mermaid
flowchart TD
    Registry[ToolRegistry]
    Registry --> Builtin[Built-in Toolkits]
    Registry --> DB[DB Tool Definitions]
    Registry --> Scanned[Scanned PHP Classes]
    Registry --> MCP[MCP Server Tools]
    DB --> Builder[Builder Tools]
    DB --> Webhook[Webhook Tools]
    DB --> RAG[RAG Tools]
```

| Source                | Type                                        | Created in                                     |
| --------------------- | ------------------------------------------- | ---------------------------------------------- |
| Built-in toolkits     | Neuron `Toolkit` classes                    | `config/neuronai-studio.php`                   |
| Builder tools         | PHP invoke body → export to class           | Studio UI (local prototyping)                  |
| Webhook tools         | HTTP endpoint + JSON schema                 | Studio UI                                      |
| RAG tools             | Knowledge base search (`KnowledgeBaseTool`) | Studio UI (`?kind=rag`)                        |
| PHP classes           | Neuron `Tool` subclasses                    | `app/Neuron/Tools/`, codegen, or package demos |
| MCP tools             | Remote MCP server                           | MCP server config                              |
| Tool Mode specialists | `NodeAsTool` via `node:{id}` refs           | Workflow canvas toolset edges                  |

Runtime resolves bindings through `ToolResolver` with prefixes such as `toolkit:`, `class:`, `tool:db:` / `db:`, `mcp:`, `provider:`, and `node:`.

### Package demo class

`IssueRefundTool` (`class:…\IssueRefundTool`) ships with the package for tool-approval demos — see the [refund-actions-agent template](/neuronai-studio/workflows/templates.md) and [Human-in-the-Loop](/neuronai-studio/workflows/human-in-the-loop.md#tool-approval). Prefer class-based tools when approval pauses must serialize across resume.

## When to use each type

| Need                                | Tool type                                                                                            |
| ----------------------------------- | ---------------------------------------------------------------------------------------------------- |
| Quick custom logic in PHP (local)   | [Builder Tool](/neuronai-studio/tools/builder-tools.md) → export class                               |
| Call an external API                | [Webhook Tool](/neuronai-studio/tools/webhook-tools.md)                                              |
| Search a knowledge base on demand   | [RAG tool](#rag-knowledge-base-tool)                                                                 |
| Reusable production class           | [Make Tool CLI](/neuronai-studio/tools/make-tool-cli.md) + export                                    |
| Math, calendar, etc.                | Built-in toolkit (`calculator`, `calendar`)                                                          |
| Filesystem, Telescope, etc.         | [MCP Server](/neuronai-studio/mcp-servers/overview.md)                                               |
| Supervisor calls a specialist agent | [Tool Mode](/neuronai-studio/workflows/ai-nodes.md#tool-mode-agent-as-tool) (`node:` / `NodeAsTool`) |

## RAG knowledge base tool

Create a tool with type `rag` to let an agent call `RagRetrievalService` during chat:

1. **Tools** → Create → kind **RAG** (or `/tools/create?kind=rag`)
2. Select `knowledge_base_id` and optional `top_k` / `threshold`
3. Bind the tool on the agent

The LLM passes a `query` string; the tool returns source-prefixed chunks. For fixed graph retrieval (not on-demand), use a [RAG workflow node](/neuronai-studio/knowledge-bases/retrieval.md) instead.

Full guide: [Knowledge Bases — Agent binding](/neuronai-studio/knowledge-bases/agent-binding.md).

## Studio routes

| Route                                        | Purpose              |
| -------------------------------------------- | -------------------- |
| `/neuronai-studio/tools`                     | List database tools  |
| `/neuronai-studio/tools/create`              | Create builder tool  |
| `/neuronai-studio/tools/create?kind=webhook` | Create webhook tool  |
| `/neuronai-studio/tools/create?kind=rag`     | Create RAG tool      |
| `/neuronai-studio/tools/{id}/edit`           | Edit tool            |
| `/neuronai-studio/tools/{id}`                | View tool details    |
| `/neuronai-studio/tools/registry?ref=...`    | Tool registry detail |

## Binding tools to agents

Tools are bound to agents by `ref` in the agent editor. Canvas toolset edges add `node:` bindings at workflow runtime. See [Creating Agents](/neuronai-studio/agents/creating-agents.md#tool-bindings) and [Canvas Editor](/neuronai-studio/workflows/canvas-editor.md#agent-tools-on-the-canvas).

## Next steps

* [Builder Tools](/neuronai-studio/tools/builder-tools.md)
* [Webhook Tools](/neuronai-studio/tools/webhook-tools.md)
* [Registry & Codegen](/neuronai-studio/tools/registry-and-codegen.md)
* [Tool Mode (AI nodes)](/neuronai-studio/workflows/ai-nodes.md#tool-mode-agent-as-tool)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://vibemx.gitbook.io/neuronai-studio/tools/overview.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.