jyapayne/langflow
1
0
Fork 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

docs: agent refactor (#8524)

Browse source 
* changes-from-1.5

* initial-content

* Apply suggestions from code review

Co-authored-by: April I. Murphy <36110273+aimurphy@users.noreply.github.com>

* breakout-tools-config

* Apply suggestions from code review

Co-authored-by: April I. Murphy <36110273+aimurphy@users.noreply.github.com>

* docs-review

* fix-broken-links

* Apply suggestions from code review

---------

Co-authored-by: April I. Murphy <36110273+aimurphy@users.noreply.github.com>
This commit is contained in:
Mendon Kissling 2025-06-20 18:05:48 -04:00 • committed by GitHub
commit 15f44519af
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
21 changed files with 572 additions and 529 deletions
Download patch file Download diff file Expand all files Collapse all files

212
docs/docs/Agents/agent-tool-calling-agent-component.md
View file

@ -1,212 +0,0 @@
---
title: Create a problem-solving agent
slug: /agents-tool-calling-agent-component
---
Developing **agents** in Langchain is complex.
The `AgentComponent` is a component for easily creating an AI agent capable of analyzing tasks using tools you provide.
The component contains all of the elements you'll need for creating an agent. Instead of managing LLM models and providers, pick your model and enter your API key. Instead of connecting a **Prompt** component, enter instructions in the component's **Agent Instruction** fields.
![Tool calling agent component](/img/tool-calling-agent-component.png)
Learn how to build a flow starting with the **Tool calling agent** component, and see how it can help you solve problems.
## Prerequisites
- [An OpenAI API key](https://platform.openai.com/)
- [A Search API key](https://www.searchapi.io/)
## Create a problem-solving agent with the Agent component
Create a problem-solving agent in Langflow, starting with the **Tool calling agent**.
1. Click **New Flow**, and then click **Blank Flow**.
2. Click and drag an **Agent** component to your workspace.
The default settings are acceptable for now, so this guide assumes you're using **Open AI** for the LLM.
3. Add your **Open AI API Key** to the **Agent** component.
4. Add **Chat input** and **Chat output** components to your flow, and connect them to the tool calling agent.
![Chat with agent component](/img/tool-calling-agent-add-chat.png)
This basic flow enables you to chat with the agent with the **Playground** after you've connected some **Tools**.
5. Connect the **Search API** tool component to your agent.
6. Add your **Search API key** to the component.
Your agent can now query the Search API for information.
7. Connect a **Calculator** tool for solving basic math problems.
8. Connect an **API Request** component to the agent.
This component is not in the **Tools** category, but the agent can still use it as a tool by enabling **Tool Mode**.
**Tool Mode** makes a component into a tool by adding a **Toolset** port that can be connected to an agent's **Tools** port.
To enable **Tool Mode** on the component, click **Tool Mode**.
The component's fields change dynamically based on the mode it's in.
![Chat with agent component](/img/tool-calling-agent-add-tools.png)
## Solve problems with the agent
Your agent now has tools for performing a web search, doing basic math, and performing API requests. You can solve many problems with just these capabilities.
* Your tabletop game group cancelled, and you're stuck at home.
Point **API Request** to an online rules document, tell your agent `You are a fun game organizer who uses the tools at your disposal`, and play a game.
* You need to learn a new software language quickly.
Point **API Request** to some docs, tell your agent `You are a knowledgeable software developer who uses the tools at your disposal`, and start learning.
See what problems you can solve with this flow. As your problem becomes more specialized, add more tools. For example, add a Python REPL component to solve math problems that are too challenging for the calculator.
### Edit a tool's metadata
To edit a tool's metadata, click the **Edit Tools** button in the tool to modify its `name`, `description`, or `enabled` metadata. These fields help connected agents understand how to use the tool, without having to modify the agent's prompt instructions.
For example, the [URL](/components-data#url) component has three tools available when **Tool Mode** is enabled.
| Tool Name | Description | Enabled |
|-----------|-------------|---------|
| `URL-fetch_content` | Use this tool to fetch and retrieve raw content from a URL, including HTML and other structured data. The full response content is returned. | true |
| `URL-fetch_content_text` | Use this tool to fetch and extract clean, readable text content from a webpage. Only plain text content is returned. | true |
| `URL-as_dataframe` | Use this tool to fetch structured data from a URL and convert it into a tabular format. Data is returned in a structured DataFrame table format. | true |
A connected agent will have a clear idea of each tool's capabilities based on the `name` and `description` metadata. The `enabled` boolean controls the tool's availability to the agent. If you think an agent is using a tool incorrectly, edit a tool's metadata to help the agent better understand the tool.
Tool names and descriptions can be edited, but the default tool identifiers cannot be changed. If you want to change the tool identifier, create a custom component.
To see which tools the agent is using and how it's using them, ask the agent, `What tools are you using to answer my questions?`
## Use an agent as a tool
The agent component itself also supports **Tool Mode** for creating multi-agent flows.
Add an agent to your problem-solving flow that uses a different OpenAI model for more specialized problem solving.
1. Click and drag an **Agent** component to your workspace.
2. Add your **Open AI API Key** to the **Agent** component.
3. In the **Model Name** field, select `gpt-4o`.
4. Click **Tool Mode** to use this new agent as a tool.
5. Connect the new agent's **Toolset** port to the previously created agent's **Tools** port.
6. Connect **Search API** and **API Request** to the new agent.
The new agent will use `gpt-4o` for the larger tasks of scraping and searching information that requires large context windows.
The problem-solving agent will now use this agent as a tool, with its unique LLM and toolset.
![The tool calling agent as a tool](/img/tool-calling-agent-as-tool.png)
7. The new agent's metadata can be edited to help the problem-solving agent understand how to use it.
Click **Edit Tools** to modify the new agent's `name` or `description` metadata so its usage is clear to the problem-solving agent.
For example, the default tool name is `Agent`. Edit the name to `Agent-gpt-4o`, and edit the description to `Use the gpt-4o model for complex problem solving`. The problem-solving agent will understand that this is the `gpt-4o` agent, and will use it for tasks requiring a larger context window.
## Add custom components as tools {#components-as-tools}
An agent can use custom components as tools.
1. To add a custom component to the problem-solving agent flow, click **New Custom Component**.
2. Add custom Python code to the custom component.
Here's an example text analyzer for sentiment analysis.
```python
from langflow.custom import Component
from langflow.io import MessageTextInput, Output
from langflow.schema import Data
import re
class TextAnalyzerComponent(Component):
display_name = "Text Analyzer"
description = "Analyzes and transforms input text."
documentation: str = "http://docs.langflow.org/components/custom"
icon = "chart-bar"
name = "TextAnalyzerComponent"
inputs = [
MessageTextInput(
name="input_text",
display_name="Input Text",
info="Enter text to analyze",
value="Hello, World!",
tool_mode=True,
),
]
outputs = [
Output(display_name="Analysis Result", name="output", method="analyze_text"),
]
def analyze_text(self) -> Data:
text = self.input_text
# Perform text analysis
word_count = len(text.split())
char_count = len(text)
sentence_count = len(re.findall(r'\w+[.!?]', text))
# Transform text
reversed_text = text[::-1]
uppercase_text = text.upper()
analysis_result = {
"original_text": text,
"word_count": word_count,
"character_count": char_count,
"sentence_count": sentence_count,
"reversed_text": reversed_text,
"uppercase_text": uppercase_text
}
data = Data(value=analysis_result)
self.status = data
return data
```
3. To enable the custom component as a tool, click **Tool Mode**.
4. Connect the tool output to the agent's tools input.
5. Ask the agent, `What tools are you using to answer my questions?`
Your response will be similar to the following, and will include your custom component.
```text
I have access to several tools that assist me in answering your questions, including:
Search API: This allows me to search for recent information or results on the web.
HTTP Requests: I can make HTTP requests to various URLs to retrieve data or interact with APIs.
Calculator: I can evaluate basic arithmetic expressions.
Text Analyzer: I can analyze and transform input text.
Current Date and Time: I can retrieve the current date and time in various time zones.
```
## Make any component a tool
If the component you want to use as a tool doesn't have a **Tool Mode** button, add `tool_mode=True` to one of the component's inputs, and connect the new **Toolset** output to the agent's **Tools** input.
Langflow supports **Tool Mode** for the following data types:
* `DataInput`
* `DataFrameInput`
* `PromptInput`
* `MessageTextInput`
* `MultilineInput`
* `DropdownInput`
For example, the [components as tools](#components-as-tools) example above adds `tool_mode=True` to the `MessageTextInput` input so the custom component can be used as a tool.
```python
inputs = [
MessageTextInput(
name="input_text",
display_name="Input Text",
info="Enter text to analyze",
value="Hello, World!",
tool_mode=True,
),
]
```
## Use the Run Flow component as a tool
An agent can use flows that are saved in your workspace as tools with the [Run flow](/components-logic#run-flow) component.
1. To add a **Run flow** component, click and drag a **Run flow** component to your workspace.
2. Select the flow you want the agent to use as a tool.
3. Enable **Tool Mode** in the component.
4. Connect the tool output to the agent's tools input.
5. To enable tool mode, select a **Flow** in the **Run flow** component, and then click **Tool Mode**.
6. Ask the agent, `What tools are you using to answer my questions?`
Your flow should be visible in the response as a tool.

14
docs/docs/Agents/agents-overview.md
View file

@ -1,14 +0,0 @@
---
title: Agents overview
slug: /agents-overview
---
**Agents** are AI systems that use LLMs as a brain to analyze problems and select external tools.
Instead of developers having to create logical statements to direct every possible path of a program, an agent can operate with autonomy. An agent can leverage external tools and APIs to gather information and take action, demonstrate chain-of-thought reasoning, and generate tailored text for specific purposes.
To simplify the development of agents, Langflow created a custom [Tool calling agent](/components-agents#agent-component) component that simplifies configuration and lets developers focus on solving problems with agents.
![Tool calling agentcomponent](/img/tool-calling-agent-component.png)
To get started, see [Create a problem solving agent](/agents-tool-calling-agent-component).

178
docs/docs/Agents/agents-tools.md Normal file
View file

@ -0,0 +1,178 @@
---
title: Configure tools for agents
slug: /agents-tools
---
import Icon from "@site/src/components/icon";
Configure tools connected to agents to extend their capabilities.
## Edit a tool component's actions
To edit a tool's actions, in the tool component, click <Icon name="SlidersHorizontal" aria-hidden="true"/> **Edit Tools** to modify its `name`, `description`, or `enabled` metadata.
These fields help connected agents understand how to use the action, without having to modify the agent's prompt instructions.
For example, the [URL](/components-data#url) component has two actions available when **Tool Mode** is enabled:
| Tool Name | Description | Enabled |
|-----------|-------------|---------|
| `fetch_content` | Fetch content from web pages recursively | true |
| `fetch_content_as_message` | Fetch web content formatted as messages | true |
A Langflow Agent has a clear idea of each tool's capabilities based on the `name` and `description` metadata. The `enabled` boolean controls the tool's availability to the agent. If you think an agent is using a tool incorrectly, edit a tool's `description` metadata to help the agent better understand the tool.
Tool names and descriptions can be edited, but the default tool identifiers cannot be changed. If you want to change the tool identifier, create a custom component.
## Use an agent as a tool
The agent component itself also supports **Tool Mode** for creating multi-agent flows.
Add an agent to your flow that uses a different OpenAI model for a larger context window.
1. Create the [Simple agent starter flow](/simple-agent).
2. Add a second agent component to the flow.
3. Add your **Open AI API Key** to the **Agent** component.
4. In the **Model Name** field, select `gpt-4.1`.
5. Click **Tool Mode** to use this new agent as a tool.
6. Connect the new agent's **Toolset** port to the previously created agent's **Tools** port.
The new agent will use `gpt-4.1` for the larger tasks of scraping and searching information that require large context windows.
The previously created agent will now use this agent as a tool, with its unique LLM and toolset.
![Agent as a tool](/img/agent-example-agent-as-tool.png)
7. The new agent's actions can be edited to help the agent understand how to use it.
Click <Icon name="SlidersHorizontal" aria-hidden="true"/> **Edit Tools** to modify its `name`, `description`, or `enabled` metadata.
For example, the default tool name is `Agent`. Edit the name to `Agent-gpt-41`, and edit the description to `Use the gpt-4.1 model for complex problem solving`. The connected agent will understand that this is the `gpt-4.1` agent, and will use it for tasks requiring a larger context window.
## Add custom components as tools {#components-as-tools}
An agent can use custom components as tools.
1. To add a custom component to the agent flow, click **New Custom Component**.
2. Add custom Python code to the custom component.
For example, to create a text analyzer component, paste the below code into the custom component's **Code** pane.
<details open>
<summary>Python</summary>
```python
from langflow.custom import Component
from langflow.io import MessageTextInput, Output
from langflow.schema import Data
import re
class TextAnalyzerComponent(Component):
display_name = "Text Analyzer"
description = "Analyzes and transforms input text."
documentation: str = "http://docs.langflow.org/components/custom"
icon = "chart-bar"
name = "TextAnalyzerComponent"
inputs = [
MessageTextInput(
name="input_text",
display_name="Input Text",
info="Enter text to analyze",
value="Hello, World!",
tool_mode=True,
),
]
outputs = [
Output(display_name="Analysis Result", name="output", method="analyze_text"),
]
def analyze_text(self) -> Data:
text = self.input_text
# Perform text analysis
word_count = len(text.split())
char_count = len(text)
sentence_count = len(re.findall(r'\w+[.!?]', text))
# Transform text
reversed_text = text[::-1]
uppercase_text = text.upper()
analysis_result = {
"original_text": text,
"word_count": word_count,
"character_count": char_count,
"sentence_count": sentence_count,
"reversed_text": reversed_text,
"uppercase_text": uppercase_text
}
data = Data(value=analysis_result)
self.status = data
return data
```
</details>
3. To use the custom component as a tool, click **Tool Mode**.
4. Connect the custom component's tool output to the agent's tools input.
5. Open the <Icon name="Play" aria-hidden="true" /> **Playground** and instruct the agent, `Use the text analyzer on this text: "Agents really are thinking machines!"`
<details open>
<summary>Response</summary>
```
AI
gpt-4o
Finished
0.6s
Here is the analysis of the text "Agents really are thinking machines!":
Original Text: Agents really are thinking machines!
Word Count: 5
Character Count: 36
Sentence Count: 1
Reversed Text: !senihcam gnikniht era yllaer stnegA
Uppercase Text: AGENTS REALLY ARE THINKING MACHINES!
```
</details>
The agent correctly calls the `analyze_text` action and returns the result to the Playground.
## Make any component a tool
If the component you want to use as a tool doesn't have a **Tool Mode** button, add `tool_mode=True` to one of the component's inputs, and connect the new **Toolset** output to the agent's **Tools** input.
Langflow supports **Tool Mode** for the following data types:
* `DataInput`
* `DataFrameInput`
* `PromptInput`
* `MessageTextInput`
* `MultilineInput`
* `DropdownInput`
For example, the [components as tools](#components-as-tools) example above adds `tool_mode=True` to the `MessageTextInput` input so the custom component can be used as a tool.
```python
inputs = [
MessageTextInput(
name="input_text",
display_name="Input Text",
info="Enter text to analyze",
value="Hello, World!",
tool_mode=True,
),
]
```
## Use flows as tools
An agent can use flows that are saved in your workspace as tools with the [Run flow](/components-logic#run-flow) component.
1. To add a **Run flow** component, click and drag a **Run flow** component to your workspace.
2. Select the flow you want the agent to use as a tool.
3. Enable **Tool Mode** in the component.
The **Run flow** component displays your flow as an available action.
4. Connect the **Run flow** component's tool output to the agent's tools input.
5. Ask the agent, `What tools are you using to answer my questions?`
Your flow should be visible in the response as a tool.
6. Ask the agent to specifically use the connected tool to answer your question.
The connected flow returns an answer based on your question.
For example, a Basic Prompting flow connected as a tool returns a different result depending upon its LLM and prompt instructions.
![Run Flow as tool connected to agnet](/img/agent-example-run-flow-as-tool.png)

91
docs/docs/Agents/agents.md Normal file
View file

@ -0,0 +1,91 @@
---
title: Use Langflow Agents
slug: /agents
---
import Icon from "@site/src/components/icon";
Agents use LLMs as a brain to autonomously analyze problems and select tools to solve them.
Langflow's [Agent component](/components-agents#agent-component) simplifies agent configuration so you can focus on application development.
The Agent component provides everything you need to create an agent, including multiple LLM providers and custom instructions.
## Agent settings
You can configure the Agent component to use your preferred provider and model, custom instructions, and tools.
### Agent models and providers
Use the **Model Provider** and **Model Name** settings to select the LLM that you want the Agent to use.
You must provide an authentication key for the selected provider, such as an OpenAI API key for OpenAI models.
### Agent instructions and input
In the **Agent Instructions** field, you can provide custom instructions that you want the Agent component to use for every conversation.
These instructions are applied in addition to the **Input**, which is provided at runtime.
### Agent tools
Agents are most useful when they have the appropriate tools available to complete requests.
An Agent component can use any Langflow component as a tool, as long as you attach it to the Agent component.
:::tip
To allow agents to use tools from MCP servers, use the [**MCP connection** component](/components-tools#mcp-connection).
:::
When you attach a component as a tool, you must configure the component as a tool by enabling **Tool Mode**.
For more information, see [Configure tools for agents](/)
## Use the Agent component in a flow
:::tip
For a pre-built demonstration, open the **Simple Agent** template flow and follow along.
:::
Create an agent in Langflow, starting with the **Agent** component and working outward.
1. Click **New Flow**, and then click **Blank Flow**.
2. Add an **Agent** component to your workspace.
3. Use the default model or select another provider and model, and then provide credentials for your chosen provider. For example, to use the default model, you must provide an OpenAI API key.
4. Add **Chat input** and **Chat output** components to your flow, and connect them to the tool calling agent.
![Chat with agent component](/img/agent-example-add-chat.png)
This basic flow allows you to chat with the agent in the **Playground**, but you're only chatting with the OpenAI LLM.
To unlock the power of the Agent component, connect some tools.
5. Add the **News Search**, **URL**, and **Calculator** components to your flow.
6. Enable **Tool Mode** in the **News Search**, **URL**, and **Calculator** components.
In the [component menu](/concepts-components#component-menu), enable **Tool mode** so you can use the component with an agent.
**Tool Mode** makes a component into a tool by modifying the component's inputs to accept requests from the Agent component to use a tool's available actions. A component in tool mode has a **Toolset** port that you must connect to an Agent component's **Tools** port if you want to allow the agent to use the tool's actions.
7. Connect the **Toolset** port on the three tool components to the **Tools** port on the Agent component.
![Chat with agent component](/img/agent-example-add-tools.png)
8. Open the <Icon name="Play" aria-hidden="true" /> **Playground**. Ask the agent, `What tools are you using to answer my questions?`
The agent should respond with a list of the connected tools.
```text
I use a combination of my built-in knowledge (up to June 2024) and a set of external tools to answer your questions. Here are the main types of tools I can use:
Web Search & Content Fetching: I can fetch and summarize content from web pages, including crawling links recursively.
News Search: I can search for recent news articles using Google News via RSS feeds.
Calculator: I can perform arithmetic calculations and evaluate mathematical expressions.
Date & Time: I can provide the current date and time in various time zones.
These tools help me provide up-to-date information, perform calculations, and retrieve specific data from the internet when needed. If you have a specific question, let me know, and I’ll use the most appropriate tool(s) to help!
```
9. Ask the agent, `Summarize today's tech news`.
The Playground displays the agent's tool calls, what input was provided, and the raw output the agent received before generating the summary. The agent should call the **News Search** component's `search_news` action.
You've successfully constructed a flow with the Langflow Agent.
Connect more tools to solve more specialized problems.
## See also
* [Configure tools for agents](/agents-tools)