API Platform: Model training in Data Controls
Owners and Admins can switch off training on inference data, which is on by default, while storage stays governed by separate settings and contracts.
Every QuiverAI update in one feed — release notes, features and changes, straight from the source.
8 updates
Owners and Admins can switch off training on inference data, which is on by default, while storage stays governed by separate settings and contracts.
QuiverAI API clients have until December 31, 2026 to move to the new x-quiverai- data posture headers before the old names are removed.
Telos is a refinement mode for demanding vector work, with a 1 million token context window, alongside the faster base Arrow 2 model.
Requests to arrow-1, arrow-1.1 and other Arrow 1.x IDs return 404 after the cutoff, so API users must switch to arrow-2 and may need new keys.
Arrow 2 speeds up SVG generation and adds a combined editing app and API platform, with Arrow 2 Telos for more complex work.
Arrow 2 models are billed by measured token usage, and new sandbox keys let developers test against the API without reaching a model.
QuiverAI's App now begins with chat, adds canvas editing and animation of SVG creations, and switches to Arrow 2 with a weekly usage meter.
Text-to-SVG generation drops 33.3% and vectorization drops 50%, alongside a higher-quality Arrow 1.1 Max variant.