Hello.
I'm Helix. What should I call you?
Get started
Skip
Artifact
‹
v1
›
Preview
Code
Select an artifact to view
Preview
Download
Done
Tap to speak
Quar
Auto
How can I help you?
Search
Settings
Language
English
Auto-speak replies
Read responses aloud via voice
Voice
Brian
▶ Test
Color mode
Dark
Aa
Font style
Serif + Sans hybrid
Haptic feedback
Web Search
Search the web to answer questions
Privacy
No accounts · Local only
Your name
Tap to change
What Helix remembers
Nothing saved yet
Clear all
Account
Sign in with Google or GitHub
AI Provider
Auto · Cloud online, Local offline
🔁 Auto
☁️ Cloud Only
🖥 Local Only
Cloud:
GPT-4.1 / GPT-4o / DeepSeek-R1 / Qwen-Coder
via Pollinations · free, no key.
Local:
Runs
in-browser
via WebGPU · full privacy · offline.
Personal API Key
Use your own OpenAI, Anthropic, Groq, Gemini, xAI, DeepSeek… key
Local AI · On-device
Not loaded — starts on first local message
Preload
Active model:
Qwen2.5-0.5B (fastest)
Kokoro AI Studio
Offline TTS · All voices · Full customization
Clear all data
Support the Project
Star Helix on GitHub · Help it grow
Support Me
Follow @Marco3113 on GitHub
Helix AI v8 · Open source · No tracking · All data stays on your device
♥
Made with love by Marco3113
Language
Done
Choose Voice
Done
TTS Language
Helix
New chat
Chats
Projects
?
You
Select Model
PRO Models Unlocked · Powered by Helix PRO
Personas
Create custom persona
Select Model
Create persona
×
Name
Accent color
Base model
GPT-4o (OpenAI)
GPT-4.1 (OpenAI)
GPT-4o mini (OpenAI)
DeepSeek R1 (Reasoning)
Qwen Coder 32B
Mistral Large
System prompt / personality
Save persona
Arena
All models
Send a prompt to see all
models respond at once
OFFLINE MODE
No internet connection.
Local AI & Kokoro Studio still work.
Dismiss
🎙️ Kokoro AI Studio
OFFLINE OK
Model
Voice
Parameters
Speed
1.0×
Pitch shift (semitones)
0
Style intensity
1.0
Volume
100%
Text to speak
0/1000
Generate & Play
Account
?
Sign out
Sign in to Helix
Sync your profile and personalize your experience
Sign in with GitHub
Generate a token at
github.com/settings/tokens
— proves you own the account
Confirm
Cancel
API Keys & Settings
Done
AI
Search
Image
TTS
Advanced
Active key
Remove
AI Provider
API Key
Model
Test
Save & Use
Keys stored only on your device · Never sent to Helix servers
Add a key for better search results. Without a key, uses DuckDuckGo free via proxy.
🦁 Brave Search API
2,000 queries/month free ·
api.search.brave.com
🔍 Serper.dev (Google)
2,500 free queries ·
serper.dev
🧭 Tavily (AI Search)
1,000 queries/month free · Optimized for AI ·
tavily.com
🐍 SerpAPI (Google/Bing/Yahoo)
100 queries/month free · Multi-engine ·
serpapi.com
⚡ Exa (Neural Search)
1,000 queries/month free · Semantic search ·
exa.ai
Choose the image generation provider. Without a key, uses FLUX.1-schnell for free.
🤗 HuggingFace Token
Free · Unlocks FLUX.1-dev ·
huggingface.co/settings/tokens
🎨 Stability AI
Stable Diffusion 3.5 · Pay-per-use ·
platform.stability.ai
⚡ fal.ai
FLUX ultra-fast · $1 free credit ·
fal.ai
Image Model
FLUX.1-dev ✨
Best quality · Requires HF token
FLUX.1-schnell
Free · Fast · No key needed
Stable Diffusion 3.5 Large 🎨
High quality · Requires Stability AI key
FLUX.1-pro (fal.ai) ⚡
Ultra fast · Requires fal.ai key
DALL-E 3 (OpenAI) 🖼
Excellent prompt following · Requires OpenAI key
Add a key to use high-quality cloud TTS voices instead of Kokoro/WebSpeech local.
TTS Provider
🔊
Kokoro / WebSpeech
Default · Free · Local · No key needed
⭐
ElevenLabs
Top quality · 10k chars/month free · Ultra-realistic voices
🤖
OpenAI TTS
Excellent · Pay-per-use · Uses your OpenAI key
🎙
Deepgram Aura
Very fast · $200 free credits · Low latency
🐟
Fish Audio
Multi-language · Voice cloning · 10k chars free
API Key
Voice ID
OpenAI Voice
${['alloy','ash','coral','echo','fable','onyx','nova','shimmer','sage'].map(v=>`
${v}
`).join('')}
OpenAI TTS Model
tts-1
Fast
tts-1-hd
High quality
4o-mini-tts
Best
Deepgram Voice
${['aura-asteria-en','aura-luna-en','aura-stella-en','aura-athena-en','aura-hera-en','aura-orion-en','aura-arcas-en','aura-perseus-en','aura-angus-en','aura-orpheus-en','aura-helios-en','aura-zeus-en'].map(v=>`
${v.replace('aura-','').replace('-en','')}
`).join('')}
Temperature
Controls randomness/creativity
0.7
Precise
Balanced
Creative
Max Response Length
Maximum tokens in each reply
4096
256
4096
8192
Custom Endpoint
For Ollama, LM Studio, or any OpenAI-compatible server
Leave blank to use the selected provider's default endpoint
System Prompt Override
Replaces the default Helix system prompt
Reset to Defaults
Loading Kokoro Voice
~80 MB · cached forever · fully local
0%
Starting download…
Cancel