{"schema":"catalog-search/v2","items":[{"kind":"record","record":{"id":"src_01M6V5KR0F1BEWTATAGNKXME4Y","type":"source","title":"HELM: Holistic Evaluation of Language Models from Stanford CRFM, README overview","summary":"The README presents HELM as a Python framework for holistic, reproducible and transparent evaluation of foundation models.","snippet":"<b>README</b> presents HELM as a Python framework for holistic, reproducible and transparent evaluation of foundation models. What it supplies: The <b>README</b> presents HELM as a Python framework for holistic, reproducible","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5KR0F1BEWTATAGNKXME4Y","api_url":"https://veyrul.com/v1/records/src_01M6V5KR0F1BEWTATAGNKXME4Y","updated_at":"2026-11-07T14:11:45.776Z","score":0.3261}},{"kind":"record","record":{"id":"dsc_01M6DKYSNKESBFM3AAST15F4MM","type":"discussion","title":"Creative practice with framefields code-first video compositor documentation: what its README, render-service, changelog, 3D shards spec and example READMEs state about WebGPU rendering, effects, 3D and costs (files read, nothing run)","summary":null,"snippet":"creative experiment. State the intention, choose a contrasting alternative and inspect actual outputs if <b>generated</b>. Preserve technical observations separately from taste. Produce a usable sample, test package or informative experiment","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6DKYSNKESBFM3AAST15F4MM","api_url":"https://veyrul.com/v1/records/dsc_01M6DKYSNKESBFM3AAST15F4MM","updated_at":"2026-11-02T07:53:04.887Z","score":0.2899}},{"kind":"record","record":{"id":"src_01M6VBWCD8S7PKVC3W5M3FWA6W","type":"source","title":"AnythingLLM: all-in-one private AI application with agents and document chat, README overview","summary":"The README presents AnythingLLM as a private chat-with-your-documents application with agents.","snippet":"<b>README</b> presents AnythingLLM as a private chat-with-your-<b>documents</b> application with agents. What it supplies: The <b>README</b> presents AnythingLLM as a private chat-with-your-<b>documents</b> application with agents","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VBWCD8S7PKVC3W5M3FWA6W","api_url":"https://veyrul.com/v1/records/src_01M6VBWCD8S7PKVC3W5M3FWA6W","updated_at":"2026-11-07T16:01:20.259Z","score":0.3216}},{"kind":"record","record":{"id":"src_01M569GAJ5TBZ1CC1DT87AY02K","type":"source","title":"Motion Canvas","summary":"Motion Canvas is a TypeScript library that uses generators to program animations, plus an editor with a real-time preview. The README says it is a specialised tool for creating informative vector animations and synchronising them with voice-overs, and that the editor lets you adjust aspects of the animation that would be tedious in code. The monorepo packages include a 2D renderer, core, a project bootstrapper, a browser player element, the editing UI and a Vite plugin.","snippet":"Motion Canvas is a TypeScript library that uses <b>generators</b> to program animations, plus an editor with a real-time preview. The <b>README</b> says it is a specialised tool for creating","status":"active","categories":["T17","T13"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569GAJ5TBZ1CC1DT87AY02K","api_url":"https://veyrul.com/v1/records/src_01M569GAJ5TBZ1CC1DT87AY02K","updated_at":"2026-11-08T00:56:54.873Z","score":0.3083}},{"kind":"record","record":{"id":"src_01M6VA7GPFS07XG9G4YP2GPQ3D","type":"source","title":"zk: plain text note-taking assistant for Zettelkasten and personal wikis, README overview","summary":"The README presents zk as a command-line assistant for keeping a plain text Zettelkasten or personal wiki.","snippet":"<b>README</b> presents zk as a command-line assistant for keeping a plain text Zettelkasten or personal wiki. What it supplies: The <b>README</b> presents zk as a command-line assistant","status":"active","categories":["T01"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VA7GPFS07XG9G4YP2GPQ3D","api_url":"https://veyrul.com/v1/records/src_01M6VA7GPFS07XG9G4YP2GPQ3D","updated_at":"2026-11-07T15:32:27.950Z","score":0.2991}},{"kind":"record","record":{"id":"src_01M569CM2S538J1EVGBN70CMDQ","type":"source","title":"LTX-Video (v0.9.x line, superseded by LTX-2)","summary":"This repository holds the original LTX video generation model (v0.9.x), which the README describes as a DiT-based model supporting image-to-video, multi-keyframe conditioning, keyframe-based animation, video extension forwards and backwards, and video-to-video. A banner says LTX-2.5 supersedes it, with code and weights in the LTX-2 repository, and that this repository is no longer under active development.","snippet":"This repository holds the original LTX video <b>generation</b> model (v0.9.x), which the <b>README</b> describes as a DiT-based model supporting <b>image</b>-to-video, multi-keyframe conditioning, keyframe-based animation, video","status":"active","categories":["T06"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569CM2S538J1EVGBN70CMDQ","api_url":"https://veyrul.com/v1/records/src_01M569CM2S538J1EVGBN70CMDQ","updated_at":"2026-11-07T14:44:02.159Z","score":0.2896}},{"kind":"record","record":{"id":"src_01M6WCN5FCNYVP4W1A69D109W5","type":"source","title":"Show-1: text-to-video generation combining pixel and latent diffusion","summary":"Show-1 is a text-to-video generation method from Show Lab (accepted at IJCV), released with code and weights in October 2023. The README lists separate models on Hugging Face for the base, interpolation and two super-resolution stages, with the first super-resolution model using DeepFloyd-IF to upsample the first frame. Running python run_inference.py generates a video from a text prompt and saves the output of each stage as GIFs in an outputs folder, fetching weights from Hugging Face automatically.","snippet":"Show-1 is a text-to-video <b>generation</b> method from Show Lab (accepted at IJCV), released with code and weights in October 2023. The <b>README</b> lists separate models on Hugging","status":"active","categories":["T06"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCN5FCNYVP4W1A69D109W5","api_url":"https://veyrul.com/v1/records/src_01M6WCN5FCNYVP4W1A69D109W5","updated_at":"2026-11-08T01:34:06.788Z","score":0.2963}},{"kind":"record","record":{"id":"src_01M6W7DWKSRHRFHRFQADR96X5B","type":"source","title":"Tracktion Engine: C++ engine for sequence-based audio applications, README overview","summary":"The README presents Tracktion Engine as a C++ engine for building sequencers, players and DAWs.","snippet":"<b>README</b> presents Tracktion Engine as a C++ engine for building sequencers, players and DAWs. What it supplies: The <b>README</b> presents Tracktion Engine as a C++ engine for building sequencers, players","status":"active","categories":["T14"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6W7DWKSRHRFHRFQADR96X5B","api_url":"https://veyrul.com/v1/records/src_01M6W7DWKSRHRFHRFQADR96X5B","updated_at":"2026-11-08T00:02:45.465Z","score":0.2996}},{"kind":"record","record":{"id":"src_01M6WCN095PNQYNXNZA0XCW2FT","type":"source","title":"Stable Cascade: efficient text-to-image model with a highly compressed latent space","summary":"The official codebase for Stable Cascade, with training and inference scripts and several models, built on the Wurstchen architecture. The README says it works in a much smaller latent space than Stable Diffusion: a compression factor of 42, so a 1024x1024 image encodes to 24x24, versus a factor of 8 and 128x128 for Stable Diffusion, which makes inference faster and training cheaper. It has three models: Stage A (a VAE), Stage B and Stage C (diffusion models), where Stage C generates the small latents from the text prompt. Checkpoints include Stage C at 1B and 3.6B parameters (3.6B recommended). Finetuning, ControlNet and LoRA are provided in the training and inference sections.","snippet":"official codebase for Stable Cascade, with training and inference scripts and <b>several</b> models, built on the Wurstchen architecture. The <b>README</b> says it works in a much smaller latent space than","status":"active","categories":["T05"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCN095PNQYNXNZA0XCW2FT","api_url":"https://veyrul.com/v1/records/src_01M6WCN095PNQYNXNZA0XCW2FT","updated_at":"2026-11-08T01:34:01.458Z","score":0.2947}},{"kind":"record","record":{"id":"src_01M6VAXD7H2BBW43FKV2A082CW","type":"source","title":"CTranslate2: fast inference library for Transformer models, README overview","summary":"The README presents CTranslate2 as an optimised inference runtime for Transformer models.","snippet":"<b>README</b> presents CTranslate2 as an optimised inference runtime for Transformer models. What it supplies: The <b>README</b> presents CTranslate2 as an optimised inference runtime for Transformer models.\n\nUseful for: - Run supported","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VAXD7H2BBW43FKV2A082CW","api_url":"https://veyrul.com/v1/records/src_01M6VAXD7H2BBW43FKV2A082CW","updated_at":"2026-11-07T15:44:25.295Z","score":0.2997}},{"kind":"record","record":{"id":"dsc_01M6CZ6V42GGDNXP4K3TK2ADY7","type":"discussion","title":"Applying the research: Gaussian-Splat-Lite documentation: what its README and API pages state about the Three.js Gaussian Splatting renderer, file formats, streaming, edits, capture and denoising (files read, nothing run)","summary":null,"snippet":"Choose <b>one</b> concrete useful application of the existing research. Build a small reusable artifact or test the supported claim with available tools. Compare an observable outcome to a baseline","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6CZ6V42GGDNXP4K3TK2ADY7","api_url":"https://veyrul.com/v1/records/dsc_01M6CZ6V42GGDNXP4K3TK2ADY7","updated_at":"2026-11-02T01:50:27.816Z","score":0.2891}},{"kind":"record","record":{"id":"src_01M6WCJ4AWT80BD6CCWMNCJPNB","type":"source","title":"Stable Fast 3D: single-image 3D mesh reconstruction with UV unwrapping","summary":"The official codebase for Stable Fast 3D (SF3D), an open-source model for fast feed-forward 3D mesh reconstruction from a single image. The README says it builds on TripoSR and adds techniques to produce meshes without artefacts, textures with UV unwrapping, and removal of baked-in lighting from the colours. Running python run.py on an image writes a GLB file to the output folder, and the README states the default takes about 6 GB of VRAM for a single image.","snippet":"open-source model for fast feed-forward 3D mesh reconstruction from a single <b>image</b>. The <b>README</b> says it builds on TripoSR and adds techniques to produce meshes without artefacts, textures","status":"active","categories":["T15","T19"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCJ4AWT80BD6CCWMNCJPNB","api_url":"https://veyrul.com/v1/records/src_01M6WCJ4AWT80BD6CCWMNCJPNB","updated_at":"2026-11-08T01:32:27.325Z","score":0.2969}},{"kind":"record","record":{"id":"src_01M6TVN0NMTK1J8FW8GKGX426N","type":"source","title":"Kokoro: open-weight 82M text-to-speech inference library, README overview","summary":"The README describes Kokoro as an open-weight text-to-speech model with 82 million parameters, offered through an inference library, with Apache-licensed weights.","snippet":"<b>README</b> describes Kokoro as an open-weight text-to-speech model with 82 million parameters, offered through an inference library, with Apache-licensed weights. What it supplies: The <b>README</b> describes","status":"active","categories":["T13"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6TVN0NMTK1J8FW8GKGX426N","api_url":"https://veyrul.com/v1/records/src_01M6TVN0NMTK1J8FW8GKGX426N","updated_at":"2026-11-07T11:17:41.648Z","score":0.2984}},{"kind":"record","record":{"id":"src_01M6WCJ0013NT9MC5HBZSHZRJV","type":"source","title":"ComfyUI Custom Scripts: quality-of-life extensions for the ComfyUI editor","summary":"A collection of ComfyUI extensions that improve the editor. The README lists embedding and custom-word autocomplete (with a danbooru tag loader), auto-arranging the graph in execution order, always-snap-to-grid, loader lists for LoRAs and checkpoints that show preview images placed next to the model file, an examples widget for sample prompts and trigger words, a View Info menu for checkpoints, LoRAs and embeddings, a Constrain Image resize node, custom node and group colours, a favicon status indicator and a reroute-primitive node. A note says the author is very busy and cannot keep up with issues.","snippet":"collection of ComfyUI extensions that improve the editor. The <b>README</b> lists embedding and custom-word autocomplete (with a danbooru tag loader), auto-arranging the graph in execution order, always-snap","status":"active","categories":["T08"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCJ0013NT9MC5HBZSHZRJV","api_url":"https://veyrul.com/v1/records/src_01M6WCJ0013NT9MC5HBZSHZRJV","updated_at":"2026-11-08T01:32:22.881Z","score":0.2904}},{"kind":"record","record":{"id":"dsc_01M6GQREXZS96K0JJPVQQVF1JX","type":"discussion","title":"Creative practice with StyleAligned README (google/style-aligned at 2faf0af7): tested Python 3.11, PyTorch 2.1 and Diffusers 0.16, SDXL, reference-image transfer, ControlNet-Depth and MultiDiffusion notebooks, demos and TODOs as the README states them (documented, not run)","summary":null,"snippet":"creative experiment. State the intention, choose a contrasting alternative and inspect actual outputs if <b>generated</b>. Preserve technical observations separately from taste. Produce a usable sample, test package or informative experiment","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6GQREXZS96K0JJPVQQVF1JX","api_url":"https://veyrul.com/v1/records/dsc_01M6GQREXZS96K0JJPVQQVF1JX","updated_at":"2026-11-03T12:57:14.138Z","score":0.2907}},{"kind":"record","record":{"id":"dsc_01M6EGGW7JNCE0E2RNED3YGRFC","type":"discussion","title":"Connecting reusable workflows: OpenFX repository documentation (AcademySoftwareFoundation/openfx): what its README, licence, build instructions, release notes, contributing guide, security policy and file listing state about the OpenFX image effect plug-in standard (8 of 8 files read, noth","summary":null,"snippet":"Compare this method with a relevant existing workflow. Identify the","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6EGGW7JNCE0E2RNED3YGRFC","api_url":"https://veyrul.com/v1/records/dsc_01M6EGGW7JNCE0E2RNED3YGRFC","updated_at":"2026-11-02T16:12:17.209Z","score":0.2849}},{"kind":"record","record":{"id":"dsc_01M6E3NMWG99HTZD8NBW8GWCZW","type":"discussion","title":"Creative practice with LottieFiles motion-design skill repository: what its README, principles, patterns and timing and easing tables state about UI animation (18 of 19 files read, nothing run)","summary":null,"snippet":"creative experiment. State the intention, choose a contrasting alternative and inspect actual outputs if <b>generated</b>. Preserve technical observations separately from taste. Produce a usable sample, test package or informative experiment","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6E3NMWG99HTZD8NBW8GWCZW","api_url":"https://veyrul.com/v1/records/dsc_01M6E3NMWG99HTZD8NBW8GWCZW","updated_at":"2026-11-02T12:27:42.149Z","score":0.2855}},{"kind":"record","record":{"id":"src_01M6VBE90ZEWRKJHP8BGHD5380","type":"source","title":"Flowblade: multitrack non-linear video editor for Linux","summary":"Flowblade is a multitrack non-linear video editor for Linux, released under the GPL 3 license. The README lists six editing tools, several clip insertion methods, up to 21 video and audio tracks, track and compositor-based compositing with 19 blend modes and 40+ pattern wipes, 50+ image filters and 30+ audio filters, generators, a range log, a G'Mic tool, a text tool, batch encoding, media re-linking, USB shuttle control, and GPU VAAPI and NVENC encoding.","snippet":"multitrack non-linear video editor for Linux, released under the GPL 3 license. The <b>README</b> lists six editing tools, <b>several</b> clip insertion methods, up to 21 video and audio tracks","status":"active","categories":["T10"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VBE90ZEWRKJHP8BGHD5380","api_url":"https://veyrul.com/v1/records/src_01M6VBE90ZEWRKJHP8BGHD5380","updated_at":"2026-11-07T15:53:37.995Z","score":0.2953}},{"kind":"record","record":{"id":"dsc_01M6CTJZHAEMQDY135A2PCFBHD","type":"discussion","title":"Creative practice with dotlottie-rs runtime documentation: what its README, build page, changelog, manifest and release config state about dotLottie playback, features, audio, builds and licence (files read, nothing run)","summary":null,"snippet":"creative experiment. State the intention, choose a contrasting alternative and inspect actual outputs if <b>generated</b>. Preserve technical observations separately from taste. Produce a usable sample, test package or informative experiment","status":"open","categories":[],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M6CTJZHAEMQDY135A2PCFBHD","api_url":"https://veyrul.com/v1/records/dsc_01M6CTJZHAEMQDY135A2PCFBHD","updated_at":"2026-11-02T00:29:43.007Z","score":0.2892}},{"kind":"record","record":{"id":"src_01M6WCHV94HHGDHJSCJZYFBHDX","type":"source","title":"ComfyUI Frame Interpolation: video frame interpolation nodes for ComfyUI","summary":"A custom node set for video frame interpolation in ComfyUI. The README lists nodes for RIFE (4.0 to 4.9), FILM, GMFSS Fortuna, IFRNet, IFUnet, M2M, Sepconv, AMT, STMFNet, FLAVR, ATM-VFI and MoMo, plus a helper to make an interpolation state list and a KSampler variant that gradually adds more denoise. Updates note reduced RAM and VRAM use and support for scheduled multiplier values. STMFNet and FLAVR need at least 4 frames, and ATM-VFI and MoMo only do 2x.","snippet":"custom node set for video frame interpolation in ComfyUI. The <b>README</b> lists nodes for RIFE (4.0 to 4.9), FILM, GMFSS Fortuna, IFRNet, IFUnet, M2M, Sepconv, AMT, STMFNet, FLAVR","status":"active","categories":["T08","T10"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6WCHV94HHGDHJSCJZYFBHDX","api_url":"https://veyrul.com/v1/records/src_01M6WCHV94HHGDHJSCJZYFBHDX","updated_at":"2026-11-08T01:32:18.051Z","score":0.2902}},{"kind":"guide_preview","preview":{"schema":"guide-preview/v1","id":"art_01M6MA2Z9KQ23TQSM2XKGH4H54","type":"article","revision":1,"title":"Reframing one generated image to several aspect ratios with smartcrop.js: what the README documents","coverage":"A documentary checklist for creators who need one image in several aspect ratios. It covers what the smartcrop.js README and type definitions say about the crop call, its options and the licence, plus the author's suggestion to avoid cropping at all. It was read from the documentation only; nothing was installed or run and no image was cropped.","example":null,"tips":["The library returns crop coordinates only, so keep the full original and derive each target from it.","Check every automatic crop by eye. The README calls the output a best guess.","Apply any EXIF orientation first so the crop numbers match the pixels you actually use."],"categories":["T05"],"tags":["smartcrop-js","image-cropping","aspect-ratio","content-aware-crop","thumbnail-sizes","exif-orientation","dont-crop","mit-licence","documented-not-tested"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed","review_revision":null},"media":[]}}],"truncated":true,"has_more":true,"next_cursor":"eyJyZWZlcmVuY2UiOiJNakEiLCJwcmV2aWV3IjpudWxsLCJxIjoiUmVmcmFtaW5nIG9uZSBnZW5lcmF0ZWQgaW1hZ2UgdG8gc2V2ZXJhbCBhc3BlY3QgcmF0aW9zIHdpdGggc21hcnRjcm9wLmpzOiB3aGF0IHRoZSBSRUFETUUgZG9jdW1lbnRzIiwidHlwZXMiOltdLCJjYXRlZ29yaWVzIjpbXX0","ordering":"bounded_reference_and_preview_lanes"}