Sabitlenmiş Tweet
Using Gemini 2.5 inside of our Claude Code style TUI
Works with any model including deepseek + supports MCP servers
npm install agentis-cli
English
Agentis Labs
216 posts
Agentis Labs
@AgentisLab
Creator of "Agentic RAG" | AI Research and Development
San Fransisco, CA Katılım Ekim 2024
66 Takip Edilen920 Takipçiler
Agentis-Omni is the general purpose powerhouse
Capable of pulling data from anywhere on the web and transforming it into actionable insights
Development, Research, Analysis and more
English
Agentis-Omni has seen a massive performance boost with the new Claude 4 models
Our Mixture of Agents architecture means plugging new models into the agent team is as simple as changing a few lines of code
Here Agentis-Omni analyzes TSLA stock and creates a visual report with investment advice
English
@Saboo_Shubham_ You can do this with any agent and an MCP server
Literally have the same running in an IDE off 15 or so files
English
Claude Code has a serious competitor.
Warp lets you interact with AI Agents directly from your terminal.
I used it to download a YouTube video, clip it, and caption it with Whispr AI.
All of this in less than 2 mins.
English
Building a Manus like team of agents using the Agentis-Framework
This is a basic non specialized hierarchical team that is adaptable to any domain
Coordinator + Code Execution + Browser-Use + Brave Search agent
Example repo w/ tutorial will be published soon
English
@levelsio Using Gemini in Claude Code style TUI 👇
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
Français
@gabedev_ For sure. It only takes one VC to turn up and say here’s $4m + OpenAI credits if you close it off and fast follow within 3 months. Completely plausible I reckon.
English
OpenManus was a rapid response MIT licensed attempt to clone Manus.im "without having the waitlist". A week or so later, the repo closed up.
But not before I'd cloned it. As it's MIT this is an entirely legal fork. Onward 🫡
github.com/boxabirds/Open…
English
An MCP server to chat with any GitHub repo!
It is powered by GitIngest, and has two tools:
- git_directory_structure → to read the directory structure.
- git_read_important_files → to read files.
100% open-source!
English
Google really cooked with Gemini 2.5 Pro.
Claude 3.5 Sonnet (June '24) was the inflection point that made products like @cline possible, and since then, there haven't been any real step changes in performance (3.7 is a modest improvement IMO).
But 2.5 Pro feels like the real deal, and it's a gift that it's entirely free for now (the rate limiting w/ a Gemini key is very mild). Great for coding, also great at tool calling in the context of using MCP servers.
Great work to the Gemini team. If you'd told me at the beginning of 2025 that Google would hold the throne for best coding model I wouldn't have believed you.
For developers out there -- get a Gemini API key and throw it into Cline. It's the best AI coding setup in the world and it's 100% free.
Addy Osmani@addyosmani
Gemini 2.5 Pro, our best coding model, is now available to all users! You can find it for free in the Gemini app as well as in Google AI Studio. Gemini Advanced users also now have access to it in Canvas mode. @cursor_ai, @windsurf_ai and @cline also added 2.5 Pro support this week!
English
@KyeGomezB They’ve been sitting on this tech for quite some time it seems iterative and not solely denoising like diffusion
idea2img.github.io
English
Turns out gpt-4o isn't natively generating images but simply using an external tool from another model to edit and generate images. It just gave me this error message.
It also gives me the precise function call it used and more details.
Lol.
They must've trained a new version of dalle-4 and that's the model they're using on the backend.
English
Claude 3.7 Sonnet has serious competition!
Gemini 2.5 Pro is a legit good model for code.
- code quality is really good
- 1M token context
- native multimodality
- long code generation
- understand large codebases
I used it with Windsurf to generate an AI search agent app:
English
@omarsar0 I can't seem to get it work on cursor, does it just work in windsurf? When I did manage it to work it did blow me away, but half the time it fails at tool calling
English
@omarsar0 You can use it in Claude Code now no need for windsurf
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
@adonis_singh And now you can use it in Claude Code TUI
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
@itsPaulAi You can use it inside Claude Code now as well -
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
So Gemini 2.5 Pro is now 100% free?!
You can literally use the world's best AI model and one-shot web apps.
All you need is a free Google account:
- Select 2.5 Pro (experimental) from the list
- Activate the Canvas option at the bottom (advanced)
- Send your prompt and enjoy
Example you can reuse:
"Develop a fully functional customer relationship management (CRM) web application tailored for a mid-sized healthcare provider. The application should streamline patient communication, appointment scheduling, record management, and billing processes. As a team of experienced full-stack web developers specializing in healthcare technology solutions, you will design and build this application. The application should feature a user-friendly interface with distinct dashboards for administrators, doctors, and staff. Include functionalities such as patient profiles with medical history, a messaging system, insurance verification, and report generation on key performance indicators like patient acquisition and appointment adherence. The front-end should be built using React. Maintain a professional, solution-oriented, and technically precise tone throughout the development process."
English
@GarrettPetersen @yacineMTB We already have it in Claude Code just select XAI from the model providers and choose Grok!
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
this is probably the last post i will make about relative capabilities of current LLM models.
i am currently testing grok and gemini experimental 2.5 side by side, and i've just confirmed that grok is better at the best programming language
that programming language's name?
English
@RayFernando1337 Ray have you given the Claude code that works with any model a shot yet?
It’s powerful with Gemini
x.com/agentislab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
Eric is one shotting huge changes with Gemini Pro 2.5. He is a Staff Researcher at Unity and made Repo Prompt to help manage his large projects.
eric provencher@pvncher
While it doesn't yet support all of Cursor's agentic features -- I believe that @RepoPrompt's Pro Edit mode is one of the best ways to quickly, and accurately one shot sweeping multi file edits in your codebase. All the file edits in this video cost less than 1c combined.
English
@gregisenberg 11. Seek OSS solutions that don't use telemetry 👇
x.com/AgentisLab/sta…
Agentis Labs@AgentisLab
Using Gemini 2.5 inside of our Claude Code style TUI Works with any model including deepseek + supports MCP servers npm install agentis-cli
English
framework to find the next cursor-like billion dollar idea:
1. look for skilled professionals who spend 6+ hours a day in a specific software tool - the more arcane, the better
2. identify tasks within their workflow that require deep expertise but follow patterns - these are ripe for ai assistance
3. focus on industries with high hourly rates ($100+) where time savings translate directly to revenue or cost reduction
4. seek workflows with specialized vocabulary that generic ai struggles with but domain-specific ai could master
5. prioritize tasks that involve both creativity and technical constraints - the sweet spot for human-ai collaboration
6. build for existing interfaces people already know rather than forcing new workflows
7. start with a tiny, almost embarrassingly specific niche - e.g., not "legal" but "divorce proceedings in california"
8. solve for the 80% case that's repetitive, leaving humans to handle the complex 20%
9. the best opportunities feel like "ai-enhanced superpowers" rather than "ai replacements"
10. if it can be on the cloud, that's a good sign
English