nousresearch

Yuanbao (元宝) groups: @mention users, query info/members.

YouTube transcripts to summaries, threads, blogs.

X/Twitter via xurl CLI: raw post search, posting, DM, media.

Create, read, edit Excel .xlsx workbooks and CSVs.

Transcribe and translate speech in 99 languages.

W&B: log ML experiments, sweeps, model registry, dashboards.

Weekly reset: commitments, stalled work, next-week plan.

Authorized web pentest: recon, proof-based exploits, report.

Poll RSS, JSON APIs, and GitHub with watermark dedup.

Unsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM.

Automate Unreal Engine editor scenes, actors, and renders.

Autonomously remove your info from data-broker sites.

TRL: SFT, DPO, GRPO, RLOO reward modeling for LLM RLHF.

Pretrain LLMs at scale with PyTorch 4D parallelism.

Drive and script tldraw offline canvases with an agent.

TDD: enforce RED-GREEN-REFACTOR, tests before code.

High-throughput LLM inference on NVIDIA GPUs.

Provision Twilio numbers, SMS/MMS, and AI outbound calls.

Teams meeting summaries, job replay, Graph subscriptions.

4-phase root cause debugging: understand bugs before fixing.

Build explorable isometric architecture atlases as HTML.

Execute plans via delegate_task subagents (2-stage review).

Provision SaaS services + sync creds via Stripe Projects.

Agent payments via Stripe Link — cards, SPT, approvals.

Stock quotes, history, search, compare, crypto via Yahoo.

Text-to-image generation, inpainting, and img2img.

Throwaway experiments to validate an idea before build.

Songwriting craft and Suno AI music prompts.

Audio spectrograms/features (mel, chroma, MFCC) via CLI.

Query Solana wallets, tokens, txs, and NFTs in USD.

Plan multi-platform social campaigns: briefs to posting.

RL post-training for LLMs with Megatron and SGLang.

Throwaway HTML mockups: 2-3 design variants to compare.

Query and edit a SiYuan knowledge base via its API.

Reference-free preference alignment, simpler than DPO.

Parallel 4-agent cleanup of recent code changes.

Rewrite text to ASD-STE100 Simplified Technical English.

Query Shopify Admin/Storefront GraphQL APIs via curl.

Shop catalog search, checkout, order tracking, returns.

Find accounts for a username across 400+ platforms.

Generate a bash wizard guiding a human through manual setup.

vLLM: high-throughput LLM serving, OpenAI API, quantization.

SAM: zero-shot image segmentation via points, boxes, masks.

Free keyless meta-search aggregating 70+ engines.

Review Kanban handoffs and route verified outcomes.

Premium scroll-driven landing pages; scroll = timeline.

Scrape sites with stealth browsing and Cloudflare bypass.

Train sparse autoencoders to interpret model features.

Read RSS, Atom, JSON feeds; discover feeds behind a page.

Debug REST/GraphQL APIs: status codes, auth, schemas, repro.

Write ML papers for NeurIPS/ICML/ICLR: design→submit.

Pre-commit review: security scan, quality gates, auto-fix.

Read Reddit: subreddits, search, threads, users. No browser.

Hybrid local search over notes, docs, and transcripts.

Vector search engine for production RAG systems.

Clean training loops with built-in distributed support.

Fully sharded data-parallel training for large models.

Debug Python: pdb REPL + debugpy remote (DAP).

Versioned site deploys to GitHub/Cloudflare/Netlify Pages.

Present property and rental listings as desktop cards.

Watch product, flight, or listing prices; alert on target.

Build creative browser demos with DOM-free text layout.

Draw code changes as animated architecture/data-flow SVGs.

Build PowerPoint decks headless with python-pptx.

Create, read, edit .pptx decks with python-pptx.

54 real design systems (Stripe, Linear, Vercel) as HTML/CSS.

Query Polymarket: markets, prices, orderbooks, history.

Play Pokemon via headless emulator + RAM reads.

Pixel art w/ era palettes (NES, Game Boy, PICO-8).

Zero-install localhost tunnels over SSH via Pinggy.

Agent RAG and long-term memory with Pinecone.

Managed vector DB for production RAG and search.

Fine-tune large LLMs with LoRA on limited GPU memory.

PDF files: create, read, merge, fill, OCR, edit text.

Agent-native web search, deep research, and enrichment.

Embed an in-page natural-language GUI copilot in web apps.

p5.js sketches: gen art, shaders, interactive, 3D.

Outlines: structured JSON/regex/Pydantic LLM generation.

GitHub supply-chain forensics: recovery, IOCs, reporting.

Follow the money via public records and sanctions data.

Control Philips Hue lights, scenes, rooms via OpenHue CLI.

Delegate coding to OpenHands CLI (model-agnostic, LiteLLM).

Delegate coding to OpenCode CLI (features, PR review).

Import an OpenClaw setup (memories, skills) into Hermes.

1-3-1 decision briefs: problem, three options, one pick.

Read, search, create, and edit notes in the Obsidian vault.

OBLITERATUS: abliterate LLM refusals (diff-in-means).

Notion API + ntn CLI: pages, databases, markdown, Workers.

Debug Node.js via --inspect + Chrome DevTools Protocol CLI.

Use live BCI cognitive and mood state from NeuroSkill.

Curate LLM training data: dedupe, filter, PII redaction.

Pay HTTP 402 APIs via Machine Payments Protocol (MPP).

Generate one- or two-ink editorial print poster images.

Serverless GPU cloud for ML jobs and model APIs.

Host modded Minecraft servers (CurseForge, Modrinth).

Build M&A accretion/dilution workbooks in Excel.

Spaced-repetition flashcards: create, review, quiz, export.

Create meme PNGs from templates with Pillow text overlay.

Turn meeting notes into cited decisions, owners, tickets.

List, auth, and call MCP servers/tools from the terminal.

Manual OAuth for remote MCP servers on headless gateways.

Geocode, POIs, routes, timezones via OpenStreetMap/OSRM.

Manim CE animations: 3Blue1Brown math/algo videos.

Karpathy's LLM Wiki: build/query interlinked markdown KB.

Vision-language chat: VQA, captioning, image dialogue.

llama.cpp local GGUF inference + HF Hub model discovery.

Build self-updating dashboards from live sources.

Build leveraged buyout workbooks with IRR/MOIC in Excel.

On-demand GPU cloud instances for ML training.

Plan and run multi-agent video production pipelines.

Iterative Python via live Jupyter kernel (hamelnb).

Design minimal cute IP mascot marks readable at 32px.

Structured LLM outputs validated with Pydantic.

Read the live Hermes desktop DOM/CSS over CDP.

Run 150+ AI apps (image, video, LLM) via inference.sh CLI.

Frontend design guidance, upstream-maintained (impeccable).

Send and receive iMessages/SMS via the imsg CLI on macOS.

Hyperliquid market data, account history, trade review.

Render MP4/WebM videos from HTML compositions.

Humanize text: strip AI-isms and add real voice.

Fast BPE/WordPiece tokenization and custom vocab training.

HuggingFace hf CLI: search/download/upload models, datasets.

Configure and troubleshoot Honcho memory for Hermes.

Himalaya CLI: IMAP/SMTP email from terminal.

Modify or debug s6 services in the Hermes Docker image.

Author in-repo SKILL.md files: frontmatter and structure.

Use, configure, theme, extend, and orchestrate Hermes Agent.

Publish sites to {slug}.here.now and store files in Drives.

HeartMuLa: Suno-like song generation from lyrics + tags.

Record a site's XHR into a HAR, derive an HTTP client.

Constrain LLM output with grammars; guarantee valid JSON.

Ground answers and documents in cited, verifiable sources.

Delegate coding to xAI Grok Build CLI (features, PRs).

Adversarial plan interview before implementation.

Gmail, Calendar, Drive, Docs, Sheets via gws CLI or Python.

Join a Google Meet call, transcribe live captions, optionally speak in realtime, and do the followup work afterwards. Use when the user asks the agent to sit in on a meeting, take notes, summarize, respond in-call, or action items from it.

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Serve an interactive codebase knowledge graph web UI.

GitHub via gh CLI: PRs, issues, reviews, repos, auth.

Search/download GIFs from Tenor via curl + jq.

Speed up long-sequence transformer training and inference.

Workout planning, macros, and body metrics via wger/USDA.

Track Apple devices/AirTags via FindMy.app on macOS.

Build, test, and deploy Python MCP servers.

Fast vector similarity search at billion scale.

Build auditable financial workbooks headless via openpyxl.

Hand-drawn Excalidraw JSON diagrams (arch, flow, seq).

Read-only EVM client: wallets, tokens, gas across 8 chains.

lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).

Triage an inbox: prioritize threads, draft replies safely.

Plan-in-code fan-outs, adversarial verification, waves.

Free keyless web, news, and image search via ddgs.

DSPy: declarative LM programs, auto-optimize prompts, RAG.

Drug discovery: ChEMBL search, drug-likeness, interactions.

Build stunning 3D scenes via a concept-art fidelity loop.

Turn a handwriting photo into an installable TTF font.

Passive recon of subdomains, SSL certs, WHOIS, and DNS.

Exploratory QA of web apps: find bugs, evidence, reports.

Create, read, edit, template, and review Word .docx files.

Extract cited obligations, deadlines, tasks from documents.

Manage Docker containers, images, volumes, and Compose.

Author/validate/export Google's DESIGN.md token spec files.

Turn an unanswerable decision into a questionnaire doc.

Build discounted cash flow valuation workbooks in Excel.

Evolve prompts/regex/SQL/code with Imbue's evolution loop.

Generate ideas via named methods from creative practice.

Generate flat, minimal educational SVG visuals as HTML.

Drive the desktop background-first; escalate on signal.

Build comparable-company valuation workbooks in Excel.

Watch named companies for material news; cited digests.

Generate images, video, and audio via diffusion workflows.

Delegate coding to OpenAI Codex CLI (features, PRs).

Inspect codebases w/ pygount: LOC, languages, ratios.

Generate wiki docs + Mermaid diagrams for any codebase.

Deploy a Worker live, no account, via wrangler --temporary.

Zero-shot image classification and image-text search.

Design one-off HTML artifacts (landing, deck, prototype).

Delegate coding to Claude Code CLI (features, PRs).

Embedding database for RAG and semantic search.

Fetch Canvas LMS courses and assignments via API token.

Box manages cloud files, sharing, search, and metadata.

Monitor blogs and RSS/Atom feeds via blogwatcher-cli tool.

Use when a fetch fails: 403/429, paywall, WAF, bot wall.

Delegate coding tasks to the Blackbox AI multi-model CLI.

Gateway to 400+ genomics and computational biology skills.

Infographics: 21 layouts x 21 styles (信息图, 可视化).

Knowledge comics (知识漫画): educational, biography, tutorial.

Article illustrations: type × style × palette consistency.

Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO).

Design and build cinematic, award-level web pages.

AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.

AST-aware structural code search and rewrite via ast-grep.

ASCII video: convert video/audio to colored ASCII MP4/GIF.

ASCII art: pyfiglet, cowsay, boxes, image-to-ascii.

Search arXiv papers by keyword, author, category, or ID.

Dark-themed SVG architecture/cloud/infra diagrams as HTML.

Validated interactive HTML diagrams, upstream-maintained.

Apple Reminders via remindctl: add, list, complete.

Manage Apple Notes via memo CLI: create, search, edit.

Operate the Antigravity CLI (agy): plugins, auth, sandbox.

Airtable REST API via curl. Records CRUD, filters, upserts.

Make a verified AI presenter video from script + image.

Use when an agent needs AgentMail CLI email inboxes.

Neutral arbiter for merge conflicts between two agents.

Roleplay a hostile user to find and triage UX pain points.

Set up Actual Computer (actual.inc) inference in Hermes.

Run PyTorch training across GPUs with minimal changes.

Build integrated IS/BS/CF financial workbooks in Excel.

Set up op CLI, sign in, and read or inject secrets.