Every AI model and tool you can use in CNAPS Studio — 80+ models and 80+ tools — each with its own page. Hosted models run on CNAPS servers, no outside account or API key needed. External models run on the provider's servers with your own API key. Use Ctrl/Cmd + F to find a model by name, or start from the quick picks below.
Not sure where to start? These are good first choices for common tasks.
| I want to… | Start with | Also try |
|---|---|---|
| Upscale a photo | PiSA-SR (sharpest) | SMFANet+, FideSR, or CoDSR for large (2K+) inputs |
| Remove a background | BiRefNet (Subject) | SAM3.1 (Scene) to pick objects by text |
| Remove an object | Moebius (fast, mask-based) | PowerPaint, LatentDiffusion |
| Edit an image with words | QWEN-2511 | FireRed-1.1 to keep a person's identity |
| Generate an image | Z-Image-Turbo | FLUX Schnell, SANA-Sprint; GLM-Image for posters with text |
| Detect objects | RF-DETR Medium | YOLO-S, DETR |
| Read documents (OCR) | GLM-OCR | DeepSeekOCR; PaddleOCR for fast CJK; Sarashina2.2-OCR for Japanese |
| Upscale a video | SeedVR2 7B (highest quality) | SeedVR2 3B, FlashVSR-v1.1 (fast) |
| Generate a video | Helios-Base | Helios-Distilled (faster), LTX-2.5 (with sound) |
| Transcribe speech | Speech Recognition (Video) | Transcribe (Gemini) |
| Turn text into speech | Qwen3-TTS | Qwen3 Voice Design, Gemini TTS |
Open-weights models that run on CNAPS Studio's own servers — no outside account or API key needed. Licenses vary by model; see each model's page.
<aside>
Generate Music with Key, Tempo & Length Controls
</aside>
<aside>
Generate Sound Effects from Text — 48 kHz, 0.5–30 s
</aside>
<aside>
Multilingual Voiceover with Preset Voices — No API Key Needed
</aside>
<aside>
Text-to-Speech in a Voice You Describe in Words
</aside>
<aside>
Transcribe Every Clip in a Video List — Timed Transcripts
</aside>
<aside>
Transcribe Speech in a Video into Timed Text
</aside>
<aside>
Flag Explicit Images Automatically
</aside>
<aside>
Classify Pedestrian Gender in Photos
</aside>
<aside>
Identify What's in a Photo (1,000 Categories)
</aside>
<aside>
Generate Images That Follow an Edge Map
</aside>
<aside>
Generate Images from Any Control Map — 6 Modes in One Model
</aside>
<aside>
Turn Any Photo into a Depth Map for ControlNet
</aside>