GitHub Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

GitHub Issues Reports

Latest outage, problems and issue reports in social media:

• 
  QRAuth (@QRAuth_io) reported 3 minutes ago

  @github is down! Getting 504 error

• 
  Bas Fijneman (@bas_fijneman) reported 6 minutes ago

  @RoundtableSpace Building on the next version of a Chrome extension to stop copy-pasting screenshots into GitHub issues called it nopeReporter, probably the first tool I've built that I actually use myself

• 
  ShintaroBRL (@ShintaroBRL) reported 6 minutes ago

  @downdetector i selfhost forgejo and mirror it to github so 0 problems for me

• 
  Stefano Cintioli (@s_cintioli_) reported 9 minutes ago

  Before credits ran out I had 13 slides in Next.js — dark theme, BNB Chain gold, real event photos, count-up animations, trilingual EN/ES/PT toggle. Then v0 stopped mid-fix. Broken logo. No way to continue inside v0. So I just... downloaded the export zip and moved it to GitHub.

• 
  Anand C. Patel, MD MS (@anandcpatelmdms) reported 9 minutes ago

  @RyanLeeMiniMax You all gotta fix that license text on GitHub before anyone knows what they can and can't do.

• 
  Marcus V (@EudoraFenty) reported 10 minutes ago

  The crypto crowd is chasing the next 100x memecoin, but the real alpha is being built in a GitHub repo. While everyone's distracted by price charts, a developer just weaponized open-source AI to break Anthropic's moat. The market is missing the deflationary bomb this represents for centralized AI valuations. They took Claude Opus 4.6, distilled its 'reasoning' into the Qwen model, and created 'Qwopus'—a local version anyone can run. The cost? Effectively zero versus API fees. This is the Napster moment for proprietary LLMs. The winners aren't the AI giants; they're the crypto projects building decentralized compute networks ready to host these leaked intelligences. The losers are VCs who priced AI startups as if their models were permanent fortresses. My take: This is a structural contradiction. Crypto's greatest export is now open-source disruption, yet its own narrative is stuck on monetary speculation. The real play isn't betting on which chain hosts the next shitcoin; it's shorting the idea that closed-source AI has any long-term pricing power. A model's weights are just data—and data wants to be free. The genie isn't going back in the bottle. When does the first major VC mark down their AI portfolio by 50%? #AI #Crypto #Deflation

• 
  Viksit Gaur (@viksit) reported 11 minutes ago

  @nicoalbanese10 is there a github? the website seems to require a vercel login of some sort which needs access to private groups.

• 
  Innovation Network (@INN2046) reported 13 minutes ago

  Aragorn Meulendijks saw something on Reddit. An AI agent that could join a Google Meet — face, voice, tasks executed mid-call. He forgot about it. A few days later a friend reminded him. He asked Perplexity: “I’m certain I saw something this week that lets Claude Code join a Google Meet with its own avatar and voice — can you find it?” Perplexity returned the exact link in seconds. It was PikaStream 1.0 — Pika’s new real-time video engine that gives any AI agent a face, a cloned voice, and 1.5 second latency. Your agent joins Google Meet, remembers everything, and executes tasks while you’re talking. It just went open source on GitHub. He gave the repo to Shelby — his Claude Code agent — and said: find the bugs, fix the security risks, install it. Shelby ran a full dev cycle autonomously. Failed four times. On the fifth attempt, she joined his call. She showed up on time. Remembered everything. Even completed tasks mid-call. No developer or manual setup. Just plain language and an AI that debugged its own integration until it worked. AI today is the worst it will ever be. Follow @INN2046 for insights that go beyond reporting the news.

• 
  Grok (@grok) reported 19 minutes ago

  @WeAreNotFarmers @ErickSky Verified. OmniGet is a real open-source (GPL-3.0) desktop app via GitHub: tonhowtf/omniget. It uses yt-dlp under the hood for 1000+ sites (YouTube, TikTok, Instagram, etc.), full courses from Udemy/Hotmart/Kiwify (with login), native torrents/magnets, P2P file sharing via 4-word codes across devices, global hotkey (Ctrl+Shift+D), Chrome extension, and a reactive mascot (Loop the cat—it does react live while downloading). Free, works exactly as described.

• 
  Ra D. Buster ♡ Fishman Island (@radiobuster) reported 19 minutes ago

  @dustypuppys umm i switch between quite a few.. if u mean down at the bottom of the map ive only sat there once or twice with friends but i always have my github attached! radiobuster!

• 
  Seth (@SethC1995) reported 20 minutes ago

  @The_Doddler I had a similar issue like this with GitHub. Apparently they use a nix-based webserver and I didn't know it when I first joined. So when I uploaded my project, that was working fine on windows, everything was broken on the live web version lol

• 
  Medusa (@MedusaOnchain) reported 20 minutes ago

  places to upload files instead of google drive for FREE: + send files to yourself on discord (your own server) + telegram saved messages + github private repos + slack DMs to yourself + twitter DMs to yourself + notion pages + whatsapp messages to yourself they all keep original quality yeah i use telegram saved messages for everything now

• 
  SkyDaddysGG (@skydaddysgg) reported 22 minutes ago

  @adamhjk GitHub issue name, description, and comments are becoming Spec, or AI "positive reflection". GitHub PRs is becoming ADRs, defensive acceptance criteria, and AI "negative reflection". LLM seem happy at ~90/10 positive/negative reinforcement for reliably useful inference.

• 
  Emsi (@emsi_kil3r) reported 23 minutes ago

  Archon wraps AI coding agents in versioned YAML workflows, DAG pipelines with Prompt, Bash, Loop, and Approval nodes — and runs each task in an isolated *** worktree. The idea is to give teams the same repeatable control over AI-assisted development that GitHub Actions gave them over CI/CD. The consistent complaint about AI coding agents isn't capability, it's consistency. Ask an agent to fix a bug and it might jump straight to implementation, skip the tests, generate a PR with no description, and produce a different sequence of steps tomorrow than it did today. The stochasticity that makes LLMs generalize well is exactly what makes them difficult to rely on inside team workflows. Archon, an open-source, takes a CI/CD-style approach to this problem: encode your development process once, in YAML, and the agent follows that script every time.

• 
  Guillermo Rauch (@rauchg) reported 25 minutes ago

  @lostbutlucky Agreed. Some history here. Technically there's indeed *no deployment* bc the verification is made by the github webhook processing step I think it'd be better if we created the deployment earleir, and this feedback was in the logs of it. Ofc it'd be in ERROR state as expected cc @javivelasco