Skip to content

feat: update Token Plan defaults and support local image inputs#114

Merged
XXPermanentXX merged 3 commits into
mainfrom
feat/token-plan-default-models
Jul 22, 2026
Merged

feat: update Token Plan defaults and support local image inputs#114
XXPermanentXX merged 3 commits into
mainfrom
feat/token-plan-default-models

Conversation

@XXPermanentXX

@XXPermanentXX XXPermanentXX commented Jul 22, 2026

Copy link
Copy Markdown
Collaborator

Summary

  • update the built-in Token Plan defaults to use qwen3.8-max-preview, wan2.7-image, and dedicated HappyHorse text-to-video, image-to-video, and reference-to-video models
  • convert local image inputs to Base64 data URIs for Token Plan image editing, vision, image-to-video, and reference-to-video commands
  • preserve the existing temporary OSS upload behavior for standard API Key profiles
  • route Wan 2.7 image generation and editing through the synchronous multimodal endpoint
  • hide full inline Base64 image content from dry-run output
  • keep Pipeline behavior unchanged

Why

Token Plan does not support the existing temporary image upload flow. Local images therefore need to be embedded directly in supported model requests, while standard API Key users must continue using the established OSS upload path.

User impact

Commands such as bl image edit --image ./image.png, bl video generate --image ./first-frame.png, bl video ref --image ./reference.png, and bl vision describe --image ./image.png now work with local images under the Token Plan profile without requiring users to upload or manually encode the image first.

Image format and size validation remains service-owned; the CLI only needs a recognizable local file extension to construct the Data URI MIME type.

Validation

  • vp check
  • Core test suite: 78 passed
  • targeted auth, config, image, video, vision, and dry-run E2E suite: 28 passed
  • verified live Token Plan text chat with qwen3.8-max-preview
  • verified live Token Plan image generation with wan2.7-image
  • verified live Token Plan local-image editing with wan2.7-image through the Base64 compatibility path
  • verified Wan 2.7 image generation and local-image editing with a standard API Key

Notes

  • Pipeline behavior is intentionally unchanged.

- switch the default text model to qwen3.8-max-preview
- add dedicated T2V, I2V, and R2V model defaults
- persist and consume per-mode video model settings
- enable thinking when validating the qwen3.8 preview model
- convert local images to Base64 for Token Plan image and video commands
- preserve the existing OSS upload flow for standard API Key profiles
- use wan2.7-image as the default image model with the sync endpoint
- hide full Base64 image content in dry-run output
- add Token Plan compatibility tests and update related docs
@XXPermanentXX
XXPermanentXX merged commit 440cbfe into main Jul 22, 2026
2 checks passed
@XXPermanentXX
XXPermanentXX deleted the feat/token-plan-default-models branch July 22, 2026 02:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant