Prompto is an open-source web application designed to make interacting with large language models (LLMs) simple and efficient.
- Switch between LLMs: Effortlessly try out different LLMs, all within your browser.
- Fine-tune responses: Adjust the temperature setting to control the creativity and risk level of the LLM's outputs.
- Chat or query: Choose between a chatbot interface that remembers conversations or a notebook for sending prompts without context.
- Save and reuse: Create templates for frequently used prompts and add context on the fly.
Prompto runs entirely within your web browser, ensuring a smooth and responsive experience. While Prompto itself doesn't require a server, the language models (LLMs) you interact with may be hosted on external servers. This means that your browser communicates directly with the chosen LLM.
For simplicity and user privacy, all settings and chat conversations are stored solely in your browser's local storage. This approach allows you to maintain your preferences and conversation history without the need for server storage.
Congrats on the launch Antonio 🚀
Is integrating with the different LLMs as easy as entering API keys? Are we required to host anything on our own server?
@jgani Thank you!
For now there are two options:
Either you use API keys for OpenAI, MistralAI, or Anthropic, which means you won't need to set up any servers, or you can run your own models locally or on a server using Ollama.
Report
massive congrats team! being able to run multiple LLMs locally and have saved prompts is a great user experience and a big time saved
a few questions:
1. I consider myself a heavy AI user and understand the difference between the LLMs, meaning I can decide which I prefer to use. however, for your 'average person', would they really care about the model for the majority of tasks they're carrying out? Who is your main target audience? Is it more targeted at super users?
2. other than the different LLMs, running locally and having saved prompts are a core feature of the product. while I think these features are great, what are your barriers to entry? it feels like ChatGPT etc. could easily implement both of these
@elliotgarreffa thank you!
There is no team, It's just me behind Prompto, working on it as a hobby and because I needed a tool that's easy to tweak for my other projects.
Regarding your points:
1) Prompto is built for everyone, from AI enthusiasts to beginners. I understand the need to make it clearer how different models benefit users. My goal is to make AI tools more accessible and simple for everyone. I'm thinking of adding more helpful guides and interactive elements to help users get the most from Prompto.
2) True, bigger companies could easily introduce features like local operations and saved prompts. But, what sets Prompto apart is its open-source spirit, letting developers customize it to suit their needs, offering a flexibility that fixed systems don't. Check out the source code and feel free to contribute: https://github.com/antsanchez/pr...
Report
💡 Bright idea
@antonio_sanchez awesome work doing this on your own. point 1, adding guides to help people know which LLMs are best to use for what use cases, would be a big selling point imo
Hey Antonio! I was working with different LLMs for many weeks and it sounds super cool. I only suggest you to apply "send" action to the enter keyboard. It's a little change but it saves a lot of time for me when I'm testing something. Anyway, happy to see you helping on this. All the best!!
@joonatanvanhala thanks!
And do not forget you can install it as PWA (see video) so that you can always access it directly from your desktop or mobile device!
I have been working on Prompto and just updated it with new features and improvements:
New Additions:
- Support for Cohere
- Chat Arena: Compare two LLMs side by side.
- Overall Refactoring: Cleaner, quicker, and smaller app.
- Bug Fixes
Key Features:
- Multi-LLM Support: APIs from Ollama, OpenAI, Anthropic, Mistral AI, and Cohere.
-Flexible Interface: Notebook for single prompts, chat UI for conversation history, save templates for frequent tasks.
- Lightweight & Offline Ready: Runs entirely in the browser, saving your data locally.
Technical Details:
Built with Angular for a smooth user experience.
Utilizes LangChain.js for direct LLM interaction within the browser or as PWA.
chatting with different large language models without leaving my browser gets an upvote from me.
I'm curious tho, with all those models to talk to, how seamless is the switchover? Does each model come with a translator for the tech-lingo-impaired among us?
Prompto
Unofficial Product Hunt Chrome Plugin
Prompto
Prompto
Build Check
Prompto
Stockle
Prompto
Prompto
Bio Calls by Cross Paths