Releases: JochenYang/opencode-vision
Releases · JochenYang/opencode-vision
Release list
v1.8.1
Release v1.8.1
Changes since v1.8.0
- Merge remote-tracking branch 'origin/main'
- chore(release): 1.8.1
- fix(vision): block native fallback
- Merge pull request #2 from Olbrasoft/feat/subagent-delegation-mode
- feat(plugin): add subagent delegation mode
v1.8.0
Release v1.8.0
Changes since v1.6.0
- chore(release): 1.8.0
- docs(readme): sync Chinese README with delegation mode docs
- feat(plugin): add subagent delegation mode
- refactor(vision): clarify external fallback role in description
- chore(release): 1.7.0
- feat(vision): inject native image capability hint in system prompt
v1.6.0
Release v1.6.0
Changes since v1.5.0
- chore(release): 1.6.0
- feat(vision): align plugin with OpenCode's internal mechanism
v1.5.0
Release v1.5.0
Changes since v1.4.0
- feat(1.5.0): P1-2 expand NATIVE_VISION (pixtral, llama, doubao, etc.), P1-5 README sync, P2-1 type assertions, P2-7 unit tests (68 pass), P2-8 install.js writes opencode.json, P2-9 uninstall cleans tmp dir, P2-10 empty path filter
v1.4.0
Release v1.4.0
Changes since v1.3.1
- fix(1.4.0): P0-2 skip compaction, P0-5 LRU cleanup, P1-1 full-base64 MD5, P1-3 write error handling, P1-4 fetch timeout, P2-2 VISION_MAX_TOKENS
v1.3.1
Release v1.3.1
Changes since v1.3.0
- fix(1.3.1): P0-1 remove dead .check files, P0-3 dedupe hint injection, P0-4 vision tool path sandbox
v1.3.0
Release v1.3.0
Changes since v1.2.0
- chore: bump to 1.3.0
- feat: skip transform for native multimodal models (gpt, claude, gemini, qwen3.5/3.6, kimi-k2.5/2.6, minimax-m3, glm-v etc)
v1.2.0
Release v1.2.0
Changes since v1.0.2
- ci: auto-delete existing release before creating new one
- ci: use commit messages as release notes body instead of github defaults
- ci: add auto-release workflow on tag push
- feat: image sequence numbering with content-hash dedup
- feat: add MiniMax VLM backend support
- 1.0.2