DeepSeek Harness plugin for LLM API rate limiting, concurrency control, FIFO queuing, token budgets, and adaptive 429 cooldown
-
Updated
Aug 20, 2026 - JavaScript
DeepSeek Harness plugin for LLM API rate limiting, concurrency control, FIFO queuing, token budgets, and adaptive 429 cooldown
FreshRSS extension: refresh feeds within the rate limits of their sites, e.g. Reddit’s one request per minute — wait instead of failing with 429 Too Many Requests, from per-site rules or rate-limit headers.
Explore Microsoft Foundry throttling, bounded retries and pacing with Python, C# / .NET 10, a local simulator UI and Bicep.
AG Auto Retry for Antigravity. Local no-patch alternative to BigInt for Agent terminated due to error + Retry cases including 500, 503, 429 and 504.
Independent guide to why youtube-transcript-api and the youtube-transcript npm package fail on cloud servers (RequestBlocked, IpBlocked, RetryError 429, TranscriptsDisabled, empty transcripts) and the free and paid fixes, with tested examples. Not affiliated with YouTube or the library maintainers.
Fix ChatGPT “This chat is unavailable” / “Unable to load this conversation” / «Не удалось загрузить этот разговор ChatGPT» errors caused by HTTP 429 / Too Many Requests. Chrome extension for conversation recovery.
To associate your repository with the http-429 topic, visit your repo's landing page and select "manage topics."