Chat
💻 Coder
🔍 Research
📰 Journalist
Vision
🤖 Voice Assistant
Image
Video
Voice
⚙ Options
0.70
Tip: our models think before answering. The reasoning is tucked into a "💭 Thinking" toggle, with the clean answer below. Use 📎 to attach a document or an image — images route the next turn through Vision (Qwen2.5-VL), then the selected chat model resumes. Click 📄 Beautiful PDF on any reply to export a typeset PDF report.
💻 Kawkav Coder · Qwen3-Coder 30B · sandbox
Language hint:
⚙ Options
0.20
Tip: try "Write a Python program that prints the first 5 prime numbers, save it as primes.py, then run it." Press Ctrl/⌘+Enter to send · Tab to indent · 📎 to share a doc.
🔍 Deep Research · multi-step web synthesis
Quick~1 min · 3 sub-questions
Standard~3 min · 5 sub-questions
Deep~6 min · 7 sub-questions
The planner decomposes your question into sub-questions, searches the web in parallel, reads the top sources, and writes a cited report. Numbered citations
[1], [2] match the References section at the bottom.
📰 Kawkav Journalist · agentic generative news
🚨 Breakingtight, fast, datelined
📝 Featurein-depth, narrative
💬 Opinionargued, voiced
🔎 Investigationmulti-source, cited
🪐 Kawkav Pressindigo + gold, vector
📷 Photojournalismdramatic reportage
🎭 Caricature3D vector scene
🎬 Cinematicfilm still, 2.39:1
📊 Infographicdata-viz poster
✍️ Customyour own prompt
🇬🇧 English
🇪🇸 Español
🇸🇦 العربية
🌐 All three
⚡ Quick~3 min · text + cover
📰 Standard~6 min · cover + infographic
🏆 Pinnacle~12 min · full multi-track
The pipeline plans an angle, runs 4 parallel research tracks (web · live Kawkav Intelligence signals · multilingual sources · ADA historical context), generates a photojournalistic cover and an animated infographic, then assembles a polished article via the flagship Qwen3 235B model.
Drag & drop an image here, or click to choose one.
⚙ Options
Generate (text → image)
Edit (upload → modify)
Qwen-Image (detailed)
FLUX.2 (fast)
⚙ Options
1024 · Square
768 · Square
512 · Square
768×1024 · Portrait
1024×768 · Landscape
12
1
2 sec
3 sec
4 sec
5 sec
6 sec
8 sec
10 sec
Drag & drop an image, or click to choose. Leave empty for text-to-video.
⚙ Options
768×512 · Landscape
512 · Square
512×768 · Portrait
640×384 · Wide
25
20
Heads-up: video takes a little while (often a minute or two; longer clips can take several minutes). The timer below shows progress. Clips include synchronized sound — press play to hear it.
Kawkav Voice — TTS lab + conversational studios
Quick TTS test
10-30 sec clip for cloning (up to 50 MB)
Voice options
0.95×
Slower = more deliberate. ES auto-defaults to 0.9.
0.50
Chatterbox emotion exaggeration.
0.80
Higher = more variation per take.
Server presets — upload overrides this.
Conversational studios
🤖 Voice Assistant — talk to a live Kawkav agent
Real-time conversation with a Kawkav AI agent through our phone network.
idle
00:00 / 03:00
Demo limited to 3 minutes. Available in English and Spanish.
Up to 3 calls per day. Allow microphone access when prompted.
Up to 3 calls per day. Allow microphone access when prompted.
Live transcript
Click Talk to Kawkav Agent to begin. The conversation will appear here.