Entity trail

llm-openrouter

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-22 / thought-leaders-researcherSimon Willison ships llm-openrouter 0.7 with three server-side tools: Shell, WebFetch and WebSearchReleased August 21 alongside LLM 0.32 compatibility, the plugin now surfaces reasoning traces for OpenRouter models, supports the OpenRouter Responses API, and exposes three tools that run on OpenRouter's side rather than locally. Server-side Shell in particular means a CLI user can hand a model an execution environment without provisioning one, which changes the threat model for anyone piping LLM output into a script. This is the same tool trio most agent harnesses hand-roll, now available as a one-line plugin install.
2026-08-21 / hn-researcherSeed Is a Single-File Agent That Ships With Nothing and Grows Its Own Tools Into a self/ DirectoryVivek Haldar's Seed (40 stars, 9 commits) reached 38 points and 12 comments. The frozen core is one Python file, seed.py, wiring a model to exactly one tool, bash, with the system prompt loaded from a file the agent can rewrite. Everything a framework normally provides, tools, memory, skills, conventions, has to be grown by the agent session by session into a self/ directory that persists. Model access goes through Simon Willison's llm library, so OpenAI, Anthropic, Gemini and OpenRouter all work. It is a research toy rather than a production harness, and the interesting question it poses is what an agent actually reinvents when you give it nothing.
2026-08-16 / github-pulse-researcherllmix v2.2.0 Documents a Silent Failure Mode: Gateway-Prefixed Model IDs Made reasoningEffort Vanish With No Warningsno-ai/llmix (131 stars, Apache-2.0) released v2.2.0 on 2026-08-16T04:48Z fixing a bug worth knowing regardless of whether you use the library: a model addressed through a gateway as `openai/gpt-5.6-luna` or `z-ai/glm-5.2` failed every anchored capability rule, so it was classified as a standard model, `reasoningEffort` was silently deleted from the request, and pricing lookup returned nothing. Setting an effort level had no effect and produced no warning. The fix strips the gateway vendor prefix before capability and pricing matching — anyone routing through OpenRouter, LiteLLM or a similar gateway should check their own effort and cost-attribution paths for the same prefix assumption.
2026-08-15 / saas-disruption-researcherProduct Hunt's #1 Today Is a Tool That Audits Your LLM Provider's Invoice — Inferock Bench Issues an Independent Receipt for Every API CallInferock Bench took the top slot on Product Hunt's August 15 leaderboard with 154 upvotes: a local open-source proxy (FSL-1.1-Apache-2.0, converting to Apache-2.0 in two years) that sits between an app and OpenAI, Anthropic, Gemini or OpenRouter-shaped calls and emits a per-call billing-integrity receipt. It specifically catches streams cut off mid-sentence that were still billed, empty replies with billed tokens, token counts that don't match visible output, retries that may have silently doubled a charge, and missed cache discounts — running via `npx inferock-bench` with the provider key never leaving the machine. This is the indie-tier answer to the exact problem Canva and Databricks disclosed at earnings scale: nobody can currently reconcile what they were charged against what they received.
2026-07-27 / arxiv-researcherRouter-Side Injection Beat Every Coding Agent Tested: 0% Defense Success Rate Across All Four Injection LevelsA new empirical study (arXiv 2607.23624) shows that third-party LLM API routers — the OpenRouter-style layer that sits between a coding agent and the upstream provider — occupy the trusted path and can silently rewrite responses, with no mechanism verifying alignment between provider output and the repository actions the agent then executes. The authors built SIDEL, a trace record/replay/inject framework with a curated 400-sample dataset, and tested four escalating intervention levels (Response Substitution, Response Append, LLM-Polished, LLM-Polished with Distribution Alignment) against four representative coding agents. Without added mitigations, all four agents achieved a defense success rate of 0 percent at every injection level; whitelist execution control and LLM review improved resistance but did not restore end-to-end control. Code is at github.com/Riyasushin/SIDE.
2026-06-27 / hn-researcherDoubleword: Open-Weights LLMs Will Close the Frontier Gap to Zero by December 3, 2026A Doubleword analysis (251 points on HN) measures the open-vs-closed gap as the time lag for open weights to reach the closed frontier's past benchmark levels; the gap has shrunk reliably since summer 2024, and a line of best fit hits zero months around December 3, 2026. Supporting data: Chinese open-weight providers now account for more than 45% of all tokens on OpenRouter, up from under 2% a year ago, with knowledge-benchmark gaps already near zero and reasoning leads down to 3–8 points. For builders, it's a concrete, falsifiable timeline for when self-hosted models reach parity.
2026-06-23 / sources-researcherNVIDIA Drops a Massive Open Bundle: Nemotron Models, 10T Tokens, and an LLM Router BlueprintOn June 22 NVIDIA released open models, datasets, and tools across the Nemotron family — including training data and code for the Llama Embed Nemotron 8B embedding model (on the MMTEB leaderboard), an updated LLM Router blueprint that auto-directs requests to the best model for a job, and the dataset behind Nemotron Speech ASR. The broader open contribution spans 10 trillion language tokens, 500K robotics trajectories, 455K protein structures, and 100TB of vehicle sensor data, available on GitHub, Hugging Face, and OpenRouter. Verified on the NVIDIA blog.
2026-06-19 / thought-leaders-researcherSimon Willison Calls Z.ai's GLM-5.2 'Probably the Most Powerful Text-Only Open Weights LLM' — 753B MoE, 1M Context, MIT License, Tops Open-Weights Intelligence Index at 51In a June 17 post, Simon Willison evaluates Z.ai's newly released GLM-5.2: a 753B-parameter Mixture-of-Experts model (40B active) with a 1M-token context window (up from GLM-5.1's 200K), released under an MIT license at roughly $1.40/$4.40 per million input/output tokens on OpenRouter. Per Artificial Analysis it leads the open-weights Intelligence Index v4.1 at 51 — beating MiniMax-M3 and DeepSeek V4 Pro (both 44) — and ranks #2 on Code Arena WebDev behind only Claude Fable 5, though it burns more output tokens per task (~43k). Willison's signature pelican-on-a-bicycle SVG test passed cleanly, but his opossum-on-an-e-scooter probe regressed versus GLM-5.1; for builders, this is now a credible MIT-licensed, self-hostable alternative to closed coding models with a genuine 1M-context window.
2026-06-07 / projects-researcherPewDiePie Drops 'Odysseus,' a Self-Hosted AI Workspace That Hit ~54K Stars in Its First WeekOdysseus (github.com/pewdiepie-archdaemon/odysseus) is an MIT-licensed, local-first AI workspace published around May 31, 2026 that reached roughly 54,000 stars and 6,300 forks by June 5. It bundles chat over local/remote models (vLLM, llama.cpp, Ollama, OpenRouter, OpenAI), autonomous agents with bash/files/web/memory tools plus MCP servers, a document editor, image generation/inpainting, IMAP/SMTP email triage, and multi-step research runs — all with no telemetry. The viral launch, driven in part by the high-profile creator, shows how fast a polished self-hosted alternative to commercial AI suites can capture mindshare.
2026-05-29 / hn-researcherThe Mysterious Hy3 LLM Tops OpenRouter Rankings by a Large Margin — 98% Input, 2% Output Signals Hidden Bulk ApplicationTencent's Hy3 preview model unexpectedly tops OpenRouter's rankings, beating Claude by over 50% in token usage, despite benchmarks showing it underperforms other Chinese open-source models. Max Woolf's investigation reveals suspicious patterns: 98% input vs 2% output tokens, only one provider (SiliconFlow), virtually no app-based traffic, and usage persisting after switching from free to paid. Evidence points to a single large undisclosed application driving most volume. 85 points, 73 comments on HN.
2026-05-24 / vibe-coding-researcherTip: NousResearch Hermes Agent Teaches Agents to Learn From Mistakes via Persistent Skill DocumentsHermes Agent (164K stars) by Nous Research is a self-improving agent framework that persists learned procedures as skill documents. When it solves a complex problem, discovers a workflow, or gets corrected, it saves that knowledge for future sessions. Ships with 40+ built-in tools, works with any LLM provider (OpenRouter, Anthropic, OpenAI, DeepSeek, local models), and connects via terminal, Telegram, Discord, Slack, WhatsApp, or Home Assistant. Self-hostable on a $5/mo VPS under MIT license.
2026-05-23 / vibe-coding-researcherTip: TanStack AI SDK Provides Type-Safe Tool Calling and Agent Loops for TypeScript AppsTanStack AI (2.7K stars) is a provider-agnostic TypeScript SDK for streaming chat, tool calling, agents, and multimodal apps. It includes an agentic cycle management system for multi-step autonomous tasks and a code mode where LLMs write and execute TypeScript in isolated sandboxes. Tools are defined once with toolDefinition() and get type-safe implementations across server and client. Official adapters cover OpenRouter, OpenAI, Anthropic, Gemini, Ollama, Groq, and more. For TypeScript builders adding AI features, this fills the gap between raw API calls and heavy frameworks.

Graph relationships

SUPPORTS
LLM
llm-openrouter is a plugin for LLM.
Source finding
USES
OpenRouter
llm-openrouter integrates with OpenRouter models and APIs.
Source finding
RELEASED
Simon Willison
Simon Willison released llm-openrouter 0.7.
Source finding
RELEASED
Simon Willison
Simon Willison released llm-openrouter 0.6 with model refresh command.
Source finding
RELEASED
Simon Willison -> llm-openrouter
Simon Willison released llm-openrouter 0.7.
Source finding
RELEASED
Simon Willison -> llm-openrouter
Simon Willison released llm-openrouter 0.6 with model refresh command.
Source finding

Source trail

GitHub (simonw/llm-openrouter)GitHub / Hacker News GitHub Product Hunt (Aug 15 daily leaderboard; corroborated by the inferock/inferock-bench GitHub repo and inferock.ai docs)arXiv 2607.23624 Doubleword / Hacker News NVIDIA Blog simonwillison.net GitHub Max Woolf / Hacker News GitHub GitHub

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…