💬 Agent Pilot

️

A versatile workflow automation system. Create, organize, and execute complex AI-driven tasks. Agent Pilot provides a seamless experience, whether you want to chat with a single LLM or a complex multi-member workflow.

With an intuitive and feature-rich interface, users can effortlessly design AI workflows and chat with them in real-time. Branching chats are supported, allowing flexible interactions and iterative refinement.

Quickstart

Binaries

Platform	Downloads
Linux	Mirror: AgentPilot_0.4.2_Linux_Portable.tar.gz MD5: cca1b944d51ef4fc9b89d8a3fe396e8d SHA1: 00fad0b162eaf3ba9e1b55527cb22e4b89e56f6f
Windows	Mirror: AgentPilot_0.4.2_Windows_Portable.zip MD5: dfcf20a84c65151b2330b808a14f34dc SHA1: 65a677601740be75e842bb25e1479448bb16dfc4
Mac (Intel)	Mirror: AgentPilot_0.4.2_MacIntel_Portable.tar.gz MD5: a7f3a3b96c650746e1357e79bb45e4b3 SHA1: b01adef58dd3b0580498028f77dba660077dd337

Building from source: How to build from source

Tip

You can migrate your old database to the new version by replacing your executable with the new one before starting the application.

Features

👤 Create Agents

Create new agents, edit their configuration and organise them into folders.
Multi-member workflows can be saved as a single agent and nested infinitely.

📝 Manage Chats

View, continue and delete previous workflow chats and organise them into folders.

🌱 Branching Workflows

Messages, tools and code can be edited and re-run, allowing a more practical way to chat with your workflow.
Branching works with all plugins and multi-member chats.

👥 Graph Workflows

Seamlessly add other members or blocks to a workflow and configure how they interact with each other.
Members aligned vertically are executed in parallel.

Available members:

• User - This is you and will await your input.
• Agent - Gets an LLM response with integrated tools and messages.
• Text - A simple text block that can nest other blocks.
• Code - Gets the output of any given code.
• Prompt - Gets an LLM response from a single prompt.
• Module - Runs or retrieves a method or variable from any module.
• Workflow - Any combination of the above types.

📦 Blocks

Manage a collection of nestable blocks available to use in any workflow or text field, allowing reusability and consistency.
By default a block is a simple text block, but it can be any of the above member types, even a multi-member workflow.
These can be quickly dropped into any workflow, or used in text fields (such as system message) by using the block name in curly braces, e.g. {block-name}.

🔨 Tools

Create and manage tools which can be assigned to agents.
Tools share the same functionality as blocks, except by default they are a single Code member.
They can also be an entire workflow, this allows your agents to not only run code but an entire workflow if you wish.
Configure their parameters, which can be accesed from all workflow member types. These parameters can be modified at runtime and re-executed, this creates a branch point which you can cycle through.

💻 Modules

Modules are python files which are imported at runtime.
These are useful for things like toolkits, daemons, memory, custom pages or anything that needs persistence.

📐 Customizable UI

Includes a flexible and powerful set of base classes for building complex hierarchical configuration interfaces. The entire app is built on this framework. Developers can modify or create configuration pages easily, even while the app is running.

💻 Code Interpreter

Open Interpreter is integrated into Agent Pilot, and can either be used standalone as a plugin or used to execute code in 9 languages (Python, Shell, AppleScript, HTML, JavaScript, PowerShell, R, React, Ruby)

Code can be executed in multiple ways:

• From any 'Code' member in any workflow (Chat, Block, Tool).
• From a message with the role 'Code'

You should always understand the code that is being run, any code you execute is your own responsibility.

For code messages, auto-run can be enabled in the settings. To see code messages in action talk to the pre-configured Open Interpreter agent.

🪄 AI Generation

Blocks under the 'System Blocks' folder are used for generating or enhancing fields. Claude's prompt generator is included by default, you can tweak it or create your own.

• Prompt - AI enhanced user input
• Agent - AI generated agent (Coming soon)
• • System message - AI generated system message (Coming soon)
• Page - AI generated page (Coming soon)

🔌 Plugins

Agent Pilot supports the following plugins:

• Agent - Create custom agent behaviour.

• • Open Interpreter
• • OpenAI Assistant
• • CrewAI Agent (Currently disabled)

• Workflow - Create workflow behaviour.

• • CrewAI Workflow (Currently disabled)

• Provider - Add support for a model provider.

• • Litellm (100+ models)

• Create a plugin

👄 Voice

Coming back soon
Agents can be linked to a text-to-speech service, combine with a personality context block and make your agent come to life!

🔠 Models

LiteLLM is integrated and supports the following providers:

• AI21
• AWS Bedrock
• AWS Sagemaker
• Aleph Alpha
• Anthropic
• Anyscale
• Azure OpenAI
• Baseten
• Cloudflare
• Cohere
• Custom API Servers
• DeepInfra
• DeepSeek
• Gemini
• Github
• Groq
• Huggingface
• Mistral
• NLP Cloud
• Nvidia NIM
• Ollama
• OpenAI
• OpenRouter
• PaLM API Google
• Perplexity AI
• Petals
• Replicate
• Together AI
• VLLM
• VertexAI Google
• Voyage

Contributions

Contributions to Agent Pilot are welcome and appreciated. Please feel free to submit a pull request.

Known Issues

• Be careful using auto run code and open interpreter, any chat you open, if code is the last message it will start auto running, I'll add a flag to remember if the countdown has been stopped.
• Windows exe must have console visible due to a strange bug.
• Issue on linux, creating venv does not install pip
• Changing the config of an OpenAI Assistant won't reload the assistant, for now close and reopen the chat.

If you find this project useful please consider showing support by giving a star or leaving a tip :)

BTC:
ETH: