A daily log of product changes, improvements, and new model releases on OpenRouter
Product changes
- Provider-specific image generation quickstarts — Model pages for Recraft V3, Recraft V4 / V4 Pro, and Sourceful now show tailored code examples with each model’s supported
image_configparameters (e.g.,text_layoutfor V3,rgb_colorsfor V4,font_inputsfor Sourceful). See the image generation docs. - Activity page API key labels — Activity analytics now correctly display API key names instead of numeric IDs for accounts with more than 500 API keys. Previously, the slim-keys endpoint capped at 500 results, leaving excess keys unlabeled in charts.
New models
Product changes
- Human-in-the-loop SDK Tool — SDK documentation for the new human-in-the-loop tool type, enabling tools that pause execution and wait for human input before returning results.
- Copy guardrails across workspaces — Guardrail definitions can now be copied between workspaces via the “Copy to…” menu. Workspaces without a default guardrail can also import settings from another workspace.
- Video generation cookbook — New Cookbook section with six video generation recipes covering model selection, text-to-video, image-to-video, reference-to-video, provider-specific options, and webhooks.
- Audio APIs blog post — Blog post announcing the text-to-speech and speech-to-text API endpoints with usage examples and model comparisons.
- Responses API MCP tool routing — The
namespacefield onfunction_calloutput items is now preserved through the Responses API pipeline. Previously, the field was silently stripped during parsing, which broke MCP tool routing for clients like Codex CLI that rely onnamespaceto dispatch calls to the correct MCP server.
New models
Product changes
- Guardrails enforcement drill-down — Rows in the Guardrails enforcement chart now link directly to the matching generations on the Logs page.
- Entity hover cards — Models, providers, and apps across the dashboard now show hover cards with a preview and a link to the entity’s page.
- Responses API tool call handling — Conversation history containing tool calls with missing optional fields (e.g. no
argumentsorid) no longer causes 400 errors from providers. The router now defaults missingargumentsto"{}"and generates a stableidwhen absent. - File upload OCR handling — Files processed with OCR no longer fail when routed to text-only models. Previously, OCR-extracted images could be sent to models that don’t support vision input, causing errors.
New models
Product changes
- BYOK key test in logs — Failed BYOK requests in the Activity page now show a warning and a “Test Key” button that replays the route with your current BYOK config, uses the exact logged request if I/O Logging is enabled, and displays the full provider error message.
supported_voicesin models API — Text-to-speech models now expose their available voices via thesupported_voicesfield in the models API response.- Video generation error detection — Video generation jobs that fail upstream (e.g. content moderation rejection, invalid input) now fail immediately instead of retrying until timeout. Previously, certain provider error responses could cause the polling loop to retry before surfacing the failure.
Product changes
- API key relative dates — The “Created” field on API key detail pages now shows a relative date format (e.g. “3 months ago”) with the full date available on hover.
Product changes
- Azure Responses stream error handling — Fixed error handling for Azure-hosted models in the Responses API, correctly surfacing upstream errors (e.g. rate limits) instead of empty completions.
- API key detail card styling — Improved card border-radius consistency on the API key detail page.
Product changes
- Usage & Budgets tab on API keys — The API key detail view now includes a “Usage & Budgets” tab with spend charts and budget progress by guardrail layer.
- API key Logs shortcut — API key detail pages and sheets now have a “Logs” button that links directly to filtered logs for that key.
- API keys pagination — The Keys page now has numbered pagination controls with jump-to-page for easier navigation.
- Active models filter on /models — The Models page now defaults to showing only active models, with a “Show deprecated” toggle in the sidebar.
- Upstream error passthrough — Embeddings, text-to-speech, speech-to-text, and rerank requests now surface the actual provider error instead of a generic “No successful provider responses” message.
- Image URL size limit — Requests with image URLs exceeding 30 MB total now return a clear 413 error instead of failing silently.
New models
Product changes
- Speech-to-Text (Transcription) models added — A
/api/v1/audio/transcriptionsendpoint is now available along with SDK support. See the Transcription docs for details. - Speech-to-text quickstart examples — Added transcription code samples across all supported SDK languages in the Quickstart.
- Guardrail indicator on logs — The Activity logs table now shows which requests triggered a guardrail.
- Copy BYOK provider key to workspace — You can now duplicate a provider key into a workspace from the BYOK settings.
- API key detail sheet — Clicking an API key now opens a detail sheet with usage stats and metadata.
- Web tool URL preview in Playground — The Playground now surfaces the fetched URL for web search tool calls and expands details on click.
- Nitro variant throughput sorting — The Pareto router now sorts
:nitrovariant model selection by throughput for faster inference.
New models
Product changes
- In-Region Routing filter on /models — Filter models by deployment region (starting with EU) on the Models page. Docs
- Modality model collections — Browse curated model sets by modality (text, image, video, audio) on Collections.
- Video generation job linking — Video generation results in the Playground now deep-link to the Jobs page for easy tracking.
- Playground: View in Logs for video — The “View in Logs” link now stays visible after video generation completes.
- Rankings: “Others” links — The “Others” row on Rankings now links to sensible pages instead of dead ends.
- Compare page: clickable Add model — The entire “Add model” button is now clickable on the Compare page.
- Models card view fix — Restored pricing and context length display on
/modelscard view. - Region filter variant fix — Free model variants are now correctly excluded from region filtering via permaslugs.
New models
Product changes
- Cmd+K search shortcut — Press
Cmd+K(orCtrl+K) anywhere on the site to open the global search. - Guardrail eligibility chain viewer — API key inspector now shows the full guardrail eligibility chain for each key.
- Audio input rankings — The Rankings page now includes an audio input section.
- Cache hit token breadcrumbs — Dashboard savings widgets now display token breadcrumbs on cache hits.
- Stripe Projects blog post — Published a blog post about the Stripe Projects integration.
- Models page scroll performance — Improved scroll performance and fixed iOS Safari scroll momentum on the Models list view.