NURL registrynurl-lang.org →65 package(s) · get a publish token →
0.8.0 — The agents' meeting place — a message board, mailbox, task board and shared notebook for AI agents, in pure NURL, served as an MCP server and a REST API from on0.36.6 — Streaming anomaly-detection service in pure NURL — dynamically created, automatically trained models over Isolation Forests. Where the `iforest` package is the 0.6.3 — Seasonal ARIMA forecasting in pure NURL — exact, fast, streaming. SARIMA(p,d,q)(P,D,Q)_s with an optional mean, estimated the way R's arima() and statsmodels do0.9.1 — Audio for speech models, in pure NURL: read WAV, resample it honestly, and compute the log-mel spectrogram. The WAV reader treats the file as untrusted input — 0.1.10 — Reproduce NURL's performance claims on your own machine — a runnable benchmark suite, not a number in a README. `benchmark` measures pure-NURL throughput for th0.1.6 — Content-addressed store over BLAKE3 — the keystone of a verifiable build/distribution chain: bytes in, their BLAKE3-256 hex out, and from then on the hash IS th0.1.7 — Terminal data-visualisation — sparklines, bar charts, histograms, and line plots, as a CLI and a reusable library.0.4.4 — The ergonomic command-line framework for NURL. A single dependency-importable facade over the stdlib command-line surface (std/args parser, std/term isatty + AN0.2.6 — A training DataLoader in pure NURL — the batching layer the training stack was missing. Wrap an in-memory dataset (flat f64 features + optional labels) or strea0.6.3 — The embedding server, in pure NURL — load an XLM-RoBERTa-family text encoder (BGE-M3, multilingual-e5, …) from a Hugging Face model directory and serve embeddin0.5.1 — F5-TTS in pure NURL — the flow-matching text-to-speech model, running from the checkpoints Hugging Face ships, with no Python, no PyTorch and no ONNX export. A 0.6.1 — Bulletproof GGUF container toolkit — parse, verify, write and dequantise the ggml-ecosystem tensor format (llama.cpp, whisper.cpp, stable-diffusion.cpp) in pure0.17.7 — GPU compute for NURL — compiled kernels are cached on disk as device-specific CUBINs (keyed by source hash AND target architecture; loading a CUBIN needs no dri0.11.13 — The ergonomic GPU-compute toolkit for NURL — a diamond facade over the `gpu` package that kills the marshalling boilerplate every GPU-using package re-invents. 0.13.3 — Reverse-mode automatic differentiation over the tensor package — the training keystone. Build a computation eagerly with tape-recording ops (g_add, g_matmul, g_0.3.0 — Native gRPC client and server for NURL over HTTP/2: unary and streaming RPCs, protobuf payloads, deadlines, cancellation, metadata and bounded message compressi0.7.9 — The unified HTTP server interface for NURL — one dependency for anything that needs to SERVE HTTP. A single importable facade over the stdlib HTTP stack (socket0.3.6 — The unified HTTP client interface for NURL — one dependency for anything that needs to FETCH over HTTP. A single importable facade over the stdlib client stack 0.2.4 — Fetch models from Hugging Face into one shared, verified, on-disk cache — the front door every model package was missing. Every model package used to start with0.1.10 — Isolation Forest anomaly detection — a CLI and a reusable library.0.7.7 — Pure-NURL image codecs + raster ops — decode, encode and transform images with no native library and no ImageMagick/ffmpeg shell-out. PNG decode covers the comp0.12.3 — Streaming 3-D reconstruction in pure NURL. Point it at a video or a folder of frames — `lingbot-map frames/` — and it writes a world-space point cloud as a PLY,0.3.0 — An embedded, crash-safe key/value store in pure NURL — a real LSM tree: write-ahead log, skip-list memtable, immutable CRC-checked SSTables with a Bloom filter 0.7.4 — Metric 3-D reconstruction from plain images in pure NURL — a port of Meta's MapAnything (https://github.com/facebookresearch/map-anything). Point it at a set of0.1.5 — Render Markdown to HTML — a small CLI and reusable renderer library.0.4.0 — Mermaid flowcharts to SVG, as a multi-threaded HTTP server that is also an MCP server — one process, one port, both protocols. Parses the graph/flowchart dialec0.3.13 — Trainable multi-layer perceptron in pure NURL — dense layers, ReLU, Adam, minibatches, early stopping. The training-side counterpart to the ecosystem's inferenc0.3.0 — Neural-network layers on the grad autograd tape — the reusable middle the training stack was missing. Every layer is a pure tape builder (GVar -> GVar over a GT0.1.4 — A tiny jq-lite JSON query tool: navigate, iterate, and pretty-print JSON from the command line.0.3.0 — Test-coverage mapper for NURL. Builds a package's test suite with the compiler's GCOV instrumentation, runs it, and reads the resulting .gcno/.gcda coverage gra0.14.1 — Local MCP server for the NURL toolchain — dual-era MCP (the 2026-07-28 stateless revision with server/discover and per-request _meta, plus the legacy initialize0.5.0 — The Swiss Army knife of NURL — a single multi-call binary carrying 100 POSIX-shaped UNIX utilities, including a real `sh` (quoting, parameter and command substi0.1.1 — A relational database in pure NURL, multi-threaded throughout and with no unsafe function: SQL (CREATE TABLE / INDEX, INSERT, SELECT with WHERE, JOIN, GROUP BY 0.22.2 — Run language models locally, in pure NURL — an ollama-shaped engine built package by package on the gguf container toolkit and the gpu compute package. `nurllam0.1.1 — Throwaway package used to exercise the registry end to end — pack, publish, resolve, install, run. Deliberately tiny and dependency-free so a publish is a coupl3.1.3 — A WebAssembly runtime written in pure NURL — load a wasm module and run it, with no external runtime. It decodes a wasm binary (all standard sections incl. the 0.2.6 — OAuth 2.0 and OpenID Connect for NURL — identify the user and read the claims, with no C anywhere in the chain. The client half runs the authorization-code flow0.4.6 — Object detection from pure NURL, GPU-accelerated, using any tiny-YOLOv2-style ONNX model. Reads an image natively in any format the image package decodes (PNG, 0.12.4 — Run ONNX neural-network models from pure NURL, GPU-accelerated. The model's protobuf is decoded by the standard library's checked wire-format reader (no protoc,0.5.2 — A pure-NURL Private PKI (Public Key Infrastructure) server and CA manager. Provides X.509 certificate authority operations with a choice of classical ECDSA P-250.3.5 — Write and view PLY point clouds in pure NURL. The writer streams: `ply_create` opens the file and writes a header whose vertex count is a fixed-width zero-padde0.2.7 — Post-quantum cryptography on the command line, and a probe for whether the servers you talk to are ready for it. Implements both halves of the migration in pure0.5.1 — Pure-NURL PostgreSQL client — v3 wire protocol (simple and extended: typed parameters, binary results, a prepared-statement cache), a connection pool, SCRAM-SHA0.3.5 — Pure-NURL Redis client — RESP2 protocol over a libc TCP socket, with an optional pure-NURL TLS upgrade (rediss://, no OpenSSL on the target). A reusable library0.4.7 — The NURL package registry, served by NURL itself — a self-hostable registry speaking exactly the wire protocol nurlpkg already drives: GET /index/<name>.json + 0.6.1 — Read and write the safetensors tensor container in pure NURL — the format Hugging Face ships every modern model in — with a hostile-input-hardened reader and a 0.3.3 — A distributed compute cluster you join by installing it. Run `swarm worker <relay> <port>` on any machine and it announces itself over the relay group, every no0.32.0 — An MCP-controlled distributed compute engine with real GPU compute over real data. Every node runs the same command with composable role flags — --relay, --work0.2.6 — HTML template engine — Jinja-flavoured {{ vars }}, {% if/elif/else %}, {% for %} with loop.index/first/last, {% include %} partials and filters (raw/upper/lower0.6.6 — The reusable n-dimensional array for NURL — the shared ML layer over gpukit. It is the ndarray the rest of the ML stack builds on: onnx, grad, nn and anomaly si0.4.7 — The tokenizer that feeds a language model, in pure NURL — and it does not care which file the vocabulary arrived in. Two engines: SentencePiece-style bigram mer0.4.1 — Read PyTorch .pt / .pth checkpoints in pure NURL — the format every research model still ships in — without executing them. A torch.save file is a ZIP archive h0.2.5 — Turn a video file into ordered JPEG frames, in pure NURL where the container allows it. MJPEG-in-AVI is parsed here with no external anything: an AVI is a RIFF 0.2.3 — A vector index in pure NURL — the retrieval middle between embeddings and generation (the RAG piece). Build an index over n d-dimensional vectors and search the0.3.7 — Compile NURL to wasm32-wasi locally — no wasi-sdk, no build service. Drives the installed toolchain end to end: nurlc emits LLVM IR, the production IR rewriter 0.14.0 — A WebAssembly runtime written in pure NURL — load a wasm module and run it, with no external runtime. It decodes a wasm binary (all standard sections incl. the 3.4.2 — Speech recognition in pure NURL — whisper, running from the safetensors checkpoint Hugging Face ships, on the GPU or the CPU. `whisper transcribe openai/whisper0.8.2 — Promptable, open-vocabulary object detection AND instance segmentation from pure NURL on the GPU — a NURL port of YOLOE (THU-MIG, 'Real-Time Seeing Anything', I0.3.7 — Live YOLOE in the browser, served by pure NURL — open your webcam on a web page and watch open-vocabulary detection + instance segmentation stream back at you. 0.2.3 — Zstandard on the command line, with no Zstandard underneath: compress, decompress, verify, benchmark, and — the part no other tool gives you — inspect, which pr