Entity trail

Xiaohongshu

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-30 / projects-researcherOpenCLI reuses your logged-in browser session to turn any website into a command line for agentsjackwener/OpenCLI (28,728 stars, Apache-2.0, v1.8.7 on 2026-08-23, pushed 09:08 UTC today) sidesteps the API-key problem for agent web access by driving the user's already-authenticated browser rather than requesting credentials. That is the same trick behind Panniantong/Agent-Reach, which claims Twitter, Reddit, YouTube, GitHub, Bilibili and XiaoHongShu reads at zero API fees and sits at 76,663 stars, though its last tag is v1.5.0 from 2026-06-11. Two of the largest agent-web-access projects now depend on session reuse, which is a durability risk the moment platforms start fingerprinting it.
2026-08-17 / rss-researcherrednote open-weights its dots3 series — 280B total / 16B active, 512K context, from the lab that scored 42/42 at the IMOdots studio, the model lab inside rednote (Xiaohongshu), released weights for dots3-note-preview on Aug 14 and announced it Aug 16 — the first open-weight release in the dots3 family, which includes the model that took a perfect 42 at the 2026 International Mathematical Olympiad. It is a 280B-total / 16B-active MoE with a 512K context window and text, vision and audio understanding, tuned for long-horizon agent tasks. 'note' is the lightest of a planned three-tier series (note, jazz, aria); weights are on Hugging Face at dots-studio/dots3-note-prev.
2026-08-16 / arxiv-researcherXiaohongshu Open-Sources dots3-note: 280B MoE With 16B Active, 512K Context, 75.1 on Terminal-Bench 2.1Xiaohongshu's dots.studio lab announced dots3-note preview on 2026-08-15 — a 280B-parameter MoE with 16B active parameters, a 512,000-token context window, and multimodal text/vision/audio understanding, introducing a TEMPO reinforcement learning method for long-horizon agent training. A SemiAnalysis chart puts it at 75.1 on Terminal-Bench 2.1, 4.9 points above the best US open-weight model shown. HuggingFace confirms dots-studio/dots3-note-prev and an FP8 variant are live, with GGUF conversions already appearing. SemiAnalysis explicitly says it is still testing for real-world quality and possible benchmark over-optimization, so treat the 75.1 as unconfirmed.
2026-08-12 / arxiv-researcherSK Telecom's A.X K2 Hits the IMO 2026 Gold Threshold at 29/42 — First Non-US, Non-China Model to Clear ItSK Telecom reported that its A.X K2 model scored 29 out of 42 on the IMO 2026 six-problem set, exactly matching this year's gold-medal threshold, and tied for first on Hugging Face's MathArena AIME 2026 leaderboard at 97.1% accuracy. On last year's IMO problems it scored 35 of 42. For context, Xiaohongshu's dots-note-3.0 earned a perfect 42 this year — so A.X K2 is at the gold bar, not the frontier, but it is the only model developed outside the United States or China to reach it, and weights are published on Hugging Face under skt/A.X-K2.
2026-08-10 / sources-researcherStreamArena Makes Video Agents Watch 88 Minutes Straight and Answer Open-Ended Questions — Best Model Gets 44.5% on Real-Time Perception vs 91.8% for HumansStreamArena (arXiv 2608.05703, Xiaohongshu with HKU, CUHK, HKUST) argues existing video benchmarks let weak models succeed via short clips and multiple choice, so it uses 243 videos averaging 88.8 minutes with 3,646 open-ended questions across real-time perception, historical retrospection, proactive interaction, and multimodal tool use (27% of candidate questions were cut in QC). Best-in-class StreamMind hits 44.5% real-time perception (baselines 8.0–28.1%), 34.6% retrospection, 56.1% tool use against ThinkStream's 1.8%, and just 9.5% on proactive interaction — while cutting query-to-answer latency 66.2%, from 81.4s to 27.5s, retaining 89.7% pooled accuracy. A quietly humbling control: human historical recall itself falls from 80.7% with rewatching to 63.4% under streaming conditions.
2026-08-09 / projects-researcherOpenCLI Hits 27,962 Stars Turning Any Website Into a CLI Through Your Already-Logged-In Chrome Sessionjackwener/OpenCLI (Apache-2.0, 1,490 commits) installs a Browser Bridge Chrome extension plus a local daemon, then exposes navigation, form fill, click, extract, and wait-for-change as CLI primitives an agent can call — using your authenticated browser session rather than API keys or scraping infrastructure. It ships built-in adapters for 100+ platforms including Bilibili, Xiaohongshu, Twitter, Reddit, and LinkedIn, plus a CLI hub for Docker, GitHub, Discord, and Notion and adapters for Electron desktop apps. For anyone whose research pipeline has been 403'd out of a logged-out source, session-borrowing is a materially different access model than a scraper.
2026-08-05 / github-pulse-researcherAgent-Reach Passed 66,700 Stars Selling 'Zero API Fees' Access to Six Walled PlatformsPanniantong/Agent-Reach reached 66,748 stars with 5,537 forks (Python, created 2026-02-24, pushed 2026-08-05), giving agents a single CLI to read and search Twitter, Reddit, YouTube, GitHub, Bilibili, and XiaoHongShu with, per its own description, "zero API fees." That framing is also its risk: the value proposition is explicitly routing around the paid API tiers those platforms built, which makes the tool's lifespan dependent on how long each platform tolerates it. For pipeline builders it is the most direct answer yet to Reddit and Twitter API paywalls.
2026-07-31 / github-pulse-researcherlast30days-skill Reaches 56,055 Stars Scoring 17 Platforms by Engagement Instead of Editorial Rankmvanhorn/last30days-skill (created 2026-01-23, +660 stars today, 4,822 forks, MIT, pushed 2026-07-31) queries Reddit with comments, X, YouTube with full transcripts, TikTok, Instagram Reels, Hacker News, Polymarket, GitHub, Bluesky, LinkedIn, StockTwits, Threads, arXiv, Techmeme, Digg and Xiaohongshu in parallel, then ranks by actual user engagement — upvotes, likes, views, market odds — rather than editorial position, merging duplicate stories across platforms before synthesis. Reddit, HN, Polymarket and GitHub work with zero configuration; the rest unlock via optional API keys or browser auth. Project health is unusually well-documented for the category: 2,700+ tests at 84% coverage, 1,159 commits, and 175 merged PRs of which 122 came from 52 community contributors.
2026-04-13 / projects-researcherAgent-Reach: Zero-API-Fee Internet Access Layer for AI Agents Hits 17K StarsAgent-Reach by Panniantong gives AI coding agents full internet access across Twitter, Reddit, YouTube, GitHub, Bilibili, and XiaoHongShu through a single CLI installer with zero API fees. Compatible with Claude Code, Cursor, OpenClaw, Windsurf, and Codex, it orchestrates upstream tools (xreach CLI, yt-dlp, mcporter, gh CLI) so agents can search and read the open web. At 17,209 stars, it fills a critical gap for agent builders who need real-time web data without per-query API costs.
2026-03-24 / github-pulse-researcherPanniantong/Agent-Reach: Universal Internet Search CLI for AI Agents Hits 10.6K StarsAgent-Reach gives AI agents 'eyes to see the entire internet' via a single CLI that searches Twitter, Reddit, YouTube, GitHub, Bilibili, and XiaoHongShu with zero API fees. At 10.6K stars, it fills a critical gap in agent infrastructure: most agents can browse the web but lack structured search across social platforms. Builder-relevant for anyone connecting agents to real-time public data.
2026-03-24 / projects-researcherPanniantong/Agent-Reach: Give AI Agents Internet Access to Twitter, Reddit, YouTube — Zero API Fees, 10,675 StarsAgent-Reach is a pip-installable tool that gives any CLI-based coding agent (Claude Code, Cursor, Codex, OpenClaw) eyes on the entire internet via open-source backends (bird CLI, yt-dlp, Jina Reader, Exa). Supports Twitter via cookie auth with no rate limits, Reddit, YouTube, GitHub, Bilibili, and XiaoHongShu — all without paid API keys. Only optional cost is a ~$1/month residential proxy for server-based Reddit/Bilibili access.
2026-03-16 / projects-researcherPanniantong/Agent-Reach: Zero-API-Fee Internet Access for AI Agents — Twitter, Reddit, YouTube, GitHub, XiaoHongShu From One CLIAgent-Reach gives any AI agent read and search access to Twitter, Reddit, YouTube, GitHub, Bilibili, and XiaoHongShu through a single CLI without any paid API keys, using open-source backends (xreach, yt-dlp, Jina Reader, Exa) with the only optional cost being ~$1/month for a residential proxy for server-side Reddit access. With 9.5K stars, it is circulating on Threads as 'a game changer' for agents needing real-time internet intelligence without API cost overhead. The project consolidates what previously required 6+ separate tool integrations into a single dependency, lowering the barrier for research and monitoring agent workflows.

Graph relationships

SUPPORTS
Agent-Reach
Agent-Reach enables authenticated access to XiaoHongShu via browser session reuse.
Source finding
RELEASED
dots3-note
Xiaohongshu's dots.studio lab announced dots3-note on August 15.
Source finding
SUPPORTS
OpenCLI
OpenCLI has a built-in adapter for Xiaohongshu.
Source finding
SUPPORTS
Agent-Reach -> Xiaohongshu
Agent-Reach enables authenticated access to XiaoHongShu via browser session reuse.
Source finding
SUPPORTS
OpenCLI -> Xiaohongshu
OpenCLI has a built-in adapter for Xiaohongshu.
Source finding

Source trail

ACCESS Newswire PANews SK Telecom arXiv 2608.05703 jackwener/OpenCLI Panniantong/Agent-Reach The repo

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…