A browser-accessible hosted demo with no public description available. Without knowing what Somnium does, it is a hard recommend — but the name and hosting context suggest a generative or creative AI experiment worth a quick look if you have two spare minutes.
Explore the radar’s memory
Every pick we’ve ever shipped, plotted as a living constellation. Picks cluster by topic; threads between clusters reveal cross-topic connections. Zoom in to expand a cluster into its picks — drag, filter, click through to the brief.
// Browse
All picks
Archive results
-
Somnium
-
FLUX.2 Klein multi-LoRA
A browser demo for running multiple LoRA adapters simultaneously on FLUX.2 Klein, letting you blend styles or concepts in a single image generation pass. Useful for ML practitioners experimenting with LoRA composition without setting up a local pipeline.
-
SHARP - 3D Gaussian Scene Prediction from Apple
Apple's SHARP model turns a single image into a 3D scene using Gaussian splatting, available as a browser demo. Worth a look if 3D asset generation is on your roadmap, though this is a research demo, not a production API.
-
FamilyLegacy
A hackathon project for capturing and preserving family stories using AI. Appears to target individuals who want to record oral histories or genealogical narratives in a structured way. Early-stage, limited production signal.
-
LTX 2.3 Distilled
A browser demo for Lightricks' LTX 2.3 video generation model, distilled for faster inference. Generates video from text or image prompts. Worth a look if you are evaluating open-weight video models for a product pipeline.
-
Jawbreaker
A hackathon-built tool with no public description available. The name and hosting context suggest a small experimental project rather than a production-ready release. Hard to evaluate without documentation.
-
LTX 2.3 Finetuned I2V
A browser demo of a finetuned LTX 2.3 image-to-video model. Upload a still image and get a short video clip back. Worth noting if you are evaluating open-weight I2V options, but there is no documentation on what the finetune actually changed.
-
Background Removal
A browser demo that strips image backgrounds using a hosted model. Does the job without an API key or account. Nothing novel here — it is the same task a dozen Canva features and open-source REMBG scripts already handle — but it is free and fast for a one-off job.
-
CommitLens
Analyzes a Git commit history and surfaces patterns — who is touching what, how often, and where the velocity is dropping. A browser demo with no install required. Useful for a technical PM who wants a quick read on a repo before a planning call.
-
ClaimReady — Claim Submission Check
A hackathon demo that checks insurance or benefits claim submissions for errors before filing. Narrow use case aimed at reducing rejection rates by catching common mistakes upfront. Prototype-grade quality — worth a look if you are building in the insurance or healthcare admin space.
-
PitchFight AI
A hackathon demo that simulates competitive pitch battles between AI-generated presentations or arguments. Rough around the edges and clearly a prototype, but the head-to-head format is a fun angle for founders who want to stress-test a pitch narrative against a simulated adversary.
-
PaddleOCR-VL-1.6 Online Demo
A browser demo of PaddleOCR-VL 1.6, a vision-language OCR model that extracts structured text from documents, tables, and mixed-layout images. Worth a quick test if you are evaluating OCR pipelines for non-English or dense-document workloads.
-
ViiTorVoice
A hosted voice demo with no description attached. Hard to evaluate without more context, but it appears to be a voice synthesis or conversion experiment — worth a thirty-second click if you are actively comparing TTS or voice-conversion tools.
-
BhasaFlow Translate — Khasi Text-to-Speech (Khasi TTS) & Machine Translation | bhasaflow.com
A browser demo for Khasi-English machine translation and text-to-speech. Khasi is a low-resource language spoken in northeast India. This fills a real gap in NLP tooling for underserved languages, though it is a demo rather than a production-ready API.
-
NSFW Z-Image-Turbo
A browser demo for uncensored image generation. No description provided, no obvious differentiator from the dozens of similar demos available. Worth noting only for completeness — this category is saturated and this entry adds nothing distinct.
-
Kokoro TTS
A browser-accessible text-to-speech demo running a model that has been turning heads for its voice quality relative to its size. If you need TTS in a product and are tired of paying ElevenLabs rates, this is the open-weight alternative worth benchmarking this week.
-
Free AI Humanizer - Bypass AI Detectors
A browser demo that rewrites AI-generated text to evade AI detection tools. No obvious differentiator over the several dozen identical tools in this space. Worth knowing exists, but this is a thin wrapper around paraphrasing with a specific evasion goal.
-
Hunyuan3D-2.0
Turn a single image or text prompt into a textured, production-ready 3D mesh without the usual Blender cleanup tax — a genuine time saver for anyone building product visualization tools, game asset pipelines, or e-commerce experiences who wants to evaluate where open 3D generation actually stands today.
-
Nano Banana PRO
A fast image generation demo space from multimodalart, apparently running a lightweight model variant. No description to go on, but multimodalart's spaces tend to be clean proofs-of-concept for new diffusion techniques. Worth a quick click if you follow the image-gen space.
-
FLUX LoRA DLC
A hosted playground for running FLUX-based image generation with a collection of swappable LoRA add-ons. Good for quickly testing style fine-tunes without standing up your own inference stack. No obvious differentiator over other FLUX spaces, but the LoRA library selection is broad.
-
OmniVoice
A hosted speech recognition and voice demo from the k2-fsa team. Appears to be a showcase for their ASR pipeline rather than a standalone product. Relevant if you are evaluating open speech recognition options for a voice feature.
-
FLUX.2 [dev]
A hosted demo space for Black Forest Labs' FLUX.2 dev image generation model. Worth bookmarking if you are evaluating image generation backends for a product, but this is a playground not a product integration — check the model weights and licensing terms before building on it.
-
FLUX.2 [Klein] 9B
A 9B-parameter image generation model from Black Forest Labs, runnable directly in-browser as a Space. Smaller footprint than earlier FLUX variants, which makes it relevant if you are self-hosting image generation and constrained on VRAM.
-
Omni Video Factory
A HuggingFace Space for AI video generation. No description provided and the branding is thin, so treat this as an early experiment rather than a production tool. Check the Space directly to see what models it wraps.
-
FireRed Image Edit 1.0 Fast
A fast image editing Space built on FireRed, letting you run instruction-based edits without a local GPU. One of many HuggingFace demos in this category; worth a quick test if you need a lightweight hosted editing endpoint.