Loosen the creative agent guidance so emoji are used only as sparse
category markers rather than forced on every list item, and tighten the
frontend emoji section-break normalization to only split before short
labelled headings.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Derive an ordered map of each inline reference image to the surrounding
prompt text describing its role, and feed it into the decision, research,
task-plan, conversation and image-prompt builders. The image prompt now
carries this placement so reference roles are not swapped.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When URL input transport is configured, the remote image API cannot fetch
localhost or private-network reference URLs. Detect such hosts and fall back
to multipart file upload so those references still reach the model.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add prompt instructions for emoji-led list formatting in visible
responses, plan narration, and image summaries, with guardrails against
hyphen chaining and empty numbered choices. Tighten the planner's
userFacingResponse description.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Send stream=true for text-to-image requests and consume the SSE
response, falling back to a non-streamed request when the stream errors
in a way that warrants it. Raises the scanner line limit to handle large
base64 image payloads.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Let the planner return a clarify/chat intent that asks 2-4 targeted
questions (imageCount 0) when essentials are missing, instead of forging
ahead. Tighten prompt guidance to preserve user wording and stated
reference-image roles, avoid inventing copy/specs, keep durable brand
constraints, and treat vague taste feedback as clarification. Also
fix the generation heartbeat goroutine to stop cleanly via defer, add
research-query fallbacks, and refresh image-only capability copy.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Give the image client a configurable ResponseHeaderTimeout (default 4m,
capped at the overall timeout) so a stalled upstream fails fast, and
retry such timeouts when an idempotency key is set so a slow-but-safe
request can recover instead of terminating. Raise the Next.js API proxy
maxDuration to 1200s to accommodate long generations.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Server: forward canvas/reference images to the image model so extract,
continue, and edit requests stay faithful to the source product.
- Add ImageRequestOptions.Images and route requests with inputs to the
/v1/images/edits endpoint, via multipart upload (default) or JSON
image_url payload, selectable through Agent.Image.InputImageTransport.
- Extract reference URLs from prompt directives, inline @image tokens,
and chat message image/mask contents; fold them into the idempotency
key and prepend a reference-fidelity instruction to the prompt.
Frontend:
- Shrink CanvasToast to a compact top pill.
- Paste an internal transparent clipboard node instead of re-uploading
when it matches the clipboard image.
- Keep agent-project polling alive on stream error while streaming.
root