Launching today

Dial
Open-source, local-first AI coding agent
3 followers
Open-source, local-first AI coding agent
3 followers
An open-source coding agent engineered for efficiency: re-reads send only the lines that changed, big files arrive summarized and reads go out in batches, so tasks use fewer tokens and finish sooner. It runs your tests before it calls a task done. Run open models on your own hardware with MLX, llama.cpp, Ollama or LM Studio. Hosted providers are optional. Built with safety, control and efficiency in mind. Desktop app and CLI for macOS, Linux and Windows.


Free
Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted
Hi Product Hunt!
Dial is a local-first AI coding agent platform designed to make AI-assisted software development more powerful, integrated, and collaborative.
It brings AI models, development tools, browser automation, desktop interaction, and multiple AI coding sessions together in one environment, with the developer's actual project at the centre.
Why Dial?
1. Local AI and model flexibility
Dial supports running AI models directly on a user's own hardware through:
Ollama
LM Studio
MLX
llama.cpp / GGUF
Users can choose local-only models, a combination of local and hosted models, or hosted models.
2. Multiple AI sessions working together
Dial is designed to support multiple coding sessions working within the same project.
Its approach includes shared claims, messages, and checks, helping coordinate work rather than treating every AI interaction as an isolated conversation.
This is particularly useful for developers who want to divide larger projects into smaller tasks and manage multiple AI workers
3. AI that can interact with more than code
Dial can:
Interact with a browser.
Click through websites.
Capture screenshots.
Collect browser performance information and HAR data.
Interact with desktop applications.
Use visual grounding to locate elements on the desktop.
This makes it more than a code-generation assistant. It aims to help with testing, debugging, and interacting with the software being developed.
5. Reduced token usage and potential cost savings
Dial also focuses on making AI coding more efficient.
It can summarize large source files and retrieves only relevant sections instead of repeatedly sending entire files to an AI model.
What we are trying to do now:
We are looking for testers! Dial is new and being revised. Looking for testers and developers to use it for real projects and report any errors or bugs as well as provide constructive feedback and ideas for improvement! Testers get early access to features as they are released and can be credited in changelogs and README if they wish to be.
Join the Discord at: https://discord.gg/DJmnppfXVM