Activity
Mon
Wed
Fri
Sun
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More

Memberships

Agent Zero

1.7k members • Free

15 contributions to Agent Zero
👀 Looking for volunteers to test Venice models 👇
Hello everyone! We're looking for our community members who actively use Agent Zero to help us test Venice.ai models and share their experience. You would be given a free API key for the time necessary. This test will take as much time as necessary. We need about 10 people. Here's how it works: 1. Comment under this post that you're interested. 2. Venice API key will be sent to you in DM (unless we're full already). There's $25 daily limit on the key, feel free to use all of it. 3. See Venice.ai models available selection here: https://docs.venice.ai/overview/models 4. Test various models for chat, utility and browser in Agent Zero with Venice provider and your API key. 5. Comment and discuss here about your experience. How to test properly: - Create a set of testing tasks that the agent can do from start to finish - Change memory folder in settings before changing chat model (otherwise it can see previous solutions of other models) - Take notes of common errors, mistakes, hallucinations - Take speed and user experience into account - Browser agent is sensitive to model quality and breaks easily, browser needs to be tested for reliability frst Let's discuss the rest here, thank you all! 👍
1 like • Sep 15
I can help with testing.
2 likes • Sep 19
@Jan Tomášek Feedback on Recent Venice AI Model Testing 1. Overall response speed is generally slow. 2. Qwen3-235B performs relatively well as a Chat Model, but as a Utility Model, the Searching memories… step takes too long. 3. Models with context support below 128k tend to run into memory issues, making them almost unusable. 4. Llama-3.3-70B is practically unusable as a Utility Model, often stuck at A0: Generating…. 5. Llama-3.1-405B works as a Chat Model, but its performance is quite poor. 6. Mistral-32-24B is not usable (404 error). Mistral-31-24B comes with built-in prompts and frequently causes JSON formatting errors. 7. Venice-Uncensored as a Chat Model with Qwen3-235B as a Utility Model works reasonably well, but Venice-Uncensored has a context limit of only 32,768, which slows down responses. 8. Overall, Qwen3-235B is the only model that provides a relatively good experience.
Chat interface design ideas 🖌️
We're collecting ideas for the chat interface redesign. Please share your ideas or sketches with us, I'll start: My idea is to design the chat like a timeline. Visual structure is ensured via typography and one vertical line. Color of the line is set by the message type. Minimizing and inner scrolling of messages would still work just like now.
Chat interface design ideas 🖌️
0 likes • Sep 11
It might also be a good idea to add some metadata on top of the existing dialog box.
🆕 The new A0 version 0.9.5 is in testing
Changelog: - Secrets management - agent can use credentials without seeing them - Agent can copy paste messages and files without rewriting them - LiteLLM global configuration field - Custom HTTP headers field for browser agent - Progressive web app support) - Extra model params support for JSON - Short IDs for files and memories to prevent LLM errors - Tunnel component frontend rework - Fix for timezone change bug - Notifications z-index fix
0 likes • Aug 31
Hi @Jan Tomášek I’ve submitted a PR that fixes a frontend rendering issue. The root cause was that drawMessageUser in messages.js created a new message element every time instead of updating the existing one. This resulted in duplicate user messages when both sendMessage() and poll() in index.js called setMessage(). I’ve updated drawMessageUser to call the shared _drawMessage function for rendering user messages, ensuring they are properly updated instead of duplicated. I’d really appreciate it if you could take a look and consider merging this PR. Thanks! https://github.com/agent0ai/agent-zero/pull/714
1 like • Sep 7
@Stirling Goetz you just need to visit your Agent Zero URL using Google Chrome or Safari. After the page finishes loading, you’ll see an install icon on the right side of the address bar. Click it and follow the steps to complete the installation.
We're looking for a PR guy/s 👨‍💼
Since we're all devs, we're looking for someone with better social skills 🙃 We don't want to search outside among people unrelated to our project. All of our current contributors come from within our community, and we would like to keep it that way. For this reason, we first ask here where we know our members. The job will probably only require an hour or two a day, but that's open to discussion. The ideal candidate should tell us what needs to be done 🙂 Also, it can be multiple people. The job will be paid, the exact compensation is to be discussed individually. A candidate should be: - fluent in English - experienced with community management or marketing, PR, etc. - passionate about Agent Zero and AI in general The work would mostly be about: - spreading the word about A0 in other communities, contacting YT channels - collecting ideas and preparing content like posts, events, maybe videos - managing the community on our socials We're not yet posting any official application forms, just let us know here what you think and if you'd be interested. Thank you all!
6 likes • Aug 25
Awesome! Agent Zero is getting better and better — it’s really exciting to see the team growing and becoming more complete.🙏
3 likes • Aug 27
@Joshua Cunningham If possible, I’d love to join as a PR guy/s — even without compensation — simply because I truly love Agent Zero. In the past, I’ve tried many AI tools such as OpenWebUI, LibreChat, SillyTavern, Khoj, Jan, and even purchased Wsty. There are countless others, but in the end, I chose Agent Zero. It’s by far the most powerful AI tool I’ve ever used, and it’s the community I genuinely want to be part of. That said, English is not my native language, and I’m not yet fluent in speaking it (though I’m working hard to improve). Because of this, I can’t really take part in video or voice calls for now, but I can fully communicate in writing. Of course, I don’t think I could be the sole PR guy/s, but I’d be happy to contribute as a backup or assistant, supporting the team in spreading and promoting Agent Zero.
Fix: Prevent auto-send when pressing Enter with a non-English input method
I’ve been frequently using non-English input methods lately and noticed that pressing Enter would accidentally send the message. This PR fixes the issue and improves the experience for users with non-English IMEs. Hope you’ll consider merging it https://github.com/agent0ai/agent-zero/pull/700
0
0
1-10 of 15
Wabi Focus
3
22points to level up
@agent-zero-4206
Agent Zero

Active 17h ago
Joined Aug 6, 2025
Powered by