详细内容
NEW✨ v0.7.3 is now live on GitHub. Check it out!
Ask Jan
The best of open-source AI in an easy-to-use product.
+4.5M downloads
Over 4 million downloads
If you are not using Jan with Jan-V1-4B, drop anything you are doing and install it now! It's a little Open gem.
BTW: yes, I'm an AC Milan fan 😢
DeepJan in action: Conducting DeepResearch at home. Local models are taking the lead.
Jan-V1 is seriously good, Apple should consider using this to power Siri. You just don’t need expensive closed models anymore for many tasks, open models have come a long way.
Jan is an amazing tool that is open source and you can deploy locally.
If you haven’t tried it already be sure to do so.
Thanks goes out to the @jandotai team!
🚨 Imagine Perplexity Pro — but open-source, private, and running fully on your machine.
✅ No API bills
✅ Real-time web search
✅ Your data stays with you
Meet jan.ai . 🔒✨
Running R1 locally on my laptop using @jandotai, asked it to imagine Nietszche is talking with Dostoyevsky and Spengler about the future of European civilization. They decide to express the core of their argument in haiku format.
Jan did a web search first, then deployed the
I really really like @jandotai
It's a very friendly app to locally run LLMs, great for privacy
I've tried others like LM Studio and Ollama and they're nice but very engineer-built, a bit too difficult for me
Jan is simple and cute and pretty and a great alternative to talk to
Jan (open source) has the potential to become a much better interface to models than anything model providers can build.
Just tried @jandotai today with @nebiusaistudio models
It's impressive, tested with @Zai_org and @Alibaba_Qwen models
Fully open source and lightweight, Web search and MCP options are also available.
well my (Astrodevil) assistant mode has some opinion on Elon vs Sama fight as
Great to see @jandotai doing their thing. They were quietly building all this time. Some of us caught on to them early on. One of the cleanest interfaces and local implementations of AI. Great Job! 👏
MLX q4 - llamacpp q4_k_m head to head with Jan-V1-4B using Jan client. Default settings.
🥇 MLX: 90 toks/s (88-96 min-max)
🥈 llama cpp + flash attn + kv q8_0 67 toks/s (64-69 min-max)
🥉 llamacpp: 49 toks/s (47-51 min-max)
Here video GGUF followed by MLX
With your Hugging Face PRO account you can now use trending models directly in one of the most beautiful native chat app ever made (Jan).
1⃣ Go to Settings -> Providers -> Hugging Face
2⃣ Click on "+" then copy/paste a model id
3⃣ Start chatting 🚀
If you are not using Jan with Jan-V1-4B, drop anything you are doing and install it now! It's a little Open gem.
BTW: yes, I'm an AC Milan fan 😢
DeepJan in action: Conducting DeepResearch at home. Local models are taking the lead.
Jan-V1 is seriously good, Apple should consider using this to power Siri. You just don’t need expensive closed models anymore for many tasks, open models have come a long way.
Jan is an amazing tool that is open source and you can deploy locally.
If you haven’t tried it already be sure to do so.
Thanks goes out to the @jandotai team!
🚨 Imagine Perplexity Pro — but open-source, private, and running fully on your machine.
✅ No API bills
✅ Real-time web search
✅ Your data stays with you
Meet jan.ai . 🔒✨
Running R1 locally on my laptop using @jandotai, asked it to imagine Nietszche is talking with Dostoyevsky and Spengler about the future of European civilization. They decide to express the core of their argument in haiku format.
Jan did a web search first, then deployed the
I really really like @jandotai
It's a very friendly app to locally run LLMs, great for privacy
I've tried others like LM Studio and Ollama and they're nice but very engineer-built, a bit too difficult for me
Jan is simple and cute and pretty and a great alternative to talk to
Jan (open source) has the potential to become a much better interface to models than anything model providers can build.
Just tried @jandotai today with @nebiusaistudio models
It's impressive, tested with @Zai_org and @Alibaba_Qwen models
Fully open source and lightweight, Web search and MCP options are also available.
well my (Astrodevil) assistant mode has some opinion on Elon vs Sama fight as
Great to see @jandotai doing their thing. They were quietly building all this time. Some of us caught on to them early on. One of the cleanest interfaces and local implementations of AI. Great Job! 👏
MLX q4 - llamacpp q4_k_m head to head with Jan-V1-4B using Jan client. Default settings.
🥇 MLX: 90 toks/s (88-96 min-max)
🥈 llama cpp + flash attn + kv q8_0 67 toks/s (64-69 min-max)
🥉 llamacpp: 49 toks/s (47-51 min-max)
Here video GGUF followed by MLX
With your Hugging Face PRO account you can now use trending models directly in one of the most beautiful native chat app ever made (Jan).
1⃣ Go to Settings -> Providers -> Hugging Face
2⃣ Click on "+" then copy/paste a model id
3⃣ Start chatting 🚀
If you are not using Jan with Jan-V1-4B, drop anything you are doing and install it now! It's a little Open gem.
BTW: yes, I'm an AC Milan fan 😢
DeepJan in action: Conducting DeepResearch at home. Local models are taking the lead.
Jan-V1 is seriously good, Apple should consider using this to power Siri. You just don’t need expensive closed models anymore for many tasks, open models have come a long way.
Jan is an amazing tool that is open source and you can deploy locally.
If you haven’t tried it already be sure to do so.
Thanks goes out to the @jandotai team!
🚨 Imagine Perplexity Pro — but open-source, private, and running fully on your machine.
✅ No API bills
✅ Real-time web search
✅ Your data stays with you
Meet jan.ai . 🔒✨
Running R1 locally on my laptop using @jandotai, asked it to imagine Nietszche is talking with Dostoyevsky and Spengler about the future of European civilization. They decide to express the core of their argument in haiku format.
Jan did a web search first, then deployed the
I really really like @jandotai
It's a very friendly app to locally run LLMs, great for privacy
I've tried others like LM Studio and Ollama and they're nice but very engineer-built, a bit too difficult for me
Jan is simple and cute and pretty and a great alternative to talk to
Jan (open source) has the potential to become a much better interface to models than anything model providers can build.
Just tried @jandotai today with @nebiusaistudio models
It's impressive, tested with @Zai_org and @Alibaba_Qwen models
Fully open source and lightweight, Web search and MCP options are also available.
well my (Astrodevil) assistant mode has some opinion on Elon vs Sama fight as
Great to see @jandotai doing their thing. They were quietly building all this time. Some of us caught on to them early on. One of the cleanest interfaces and local implementations of AI. Great Job! 👏
MLX q4 - llamacpp q4_k_m head to head with Jan-V1-4B using Jan client. Default settings.
🥇 MLX: 90 toks/s (88-96 min-max)
🥈 llama cpp + flash attn + kv q8_0 67 toks/s (64-69 min-max)
🥉 llamacpp: 49 toks/s (47-51 min-max)
Here video GGUF followed by MLX
With your Hugging Face PRO account you can now use trending models directly in one of the most beautiful native chat app ever made (Jan).
1⃣ Go to Settings -> Providers -> Hugging Face
2⃣ Click on "+" then copy/paste a model id
3⃣ Start chatting 🚀
Towards Open Superintelligence
Jan takes the best of open source AI and packages it into an easy-to-use product.
Github39.5K stars Discord1.8KOnline 
HuggingFace71 models
Best of open-source AI in one app
1
Models
Choose from open models or plug in your favorite online models.
ChatGPTOpenAI
ClaudeAnthropic
GeminiGoogle
LlamaMeta
MistralMistral AI
QwenAlibaba
DeepSeekDeepSeek
GemmaGoogle
KimiMoonshot AI
2
Connectors
Connect your email, files, notes and calendar. Jan works where you work.
GmailOrganize your inbox
AmazonShop for products
GoogleSearch the web
NotionWrite and organize
FigmaDesign with AI
YouTubeLook for videos
SlackRead channel messages
Google DriveFind and fetch files
JiraManage tickets
3
MemoryComing Soon
Your context carries over, so you don’t repeat yourself. Jan remembers your context and preferences.
Joe
Designer, Singapore
Things Jan keeps in mind
- •Minimalist UI tasted
- •Currently on a portfolio refresh
- •Wants brief, to-the-point answers
- •Frequent Figma/prototyping questions
- •Dark-mode sharer
- •Curious about type trends (Mostly harmless)
Built in Public
Our core team believes that AI should be open,
and Jan is built in public.
Develop
Submit PRs for UI, tooling, or edge optimizations.
Share
Spread the word, write tutorials, host meet-ups, post videos.
Train
Add evals, safety tests, or training recipes.
Ask Jan, get things done
+4.5M downloads, Free & Open source