Peter Steinberger
69980e8bf4
fix: resolve bare model ids via allowlist ( #51580 ) (thanks @honwee)
2026-04-04 17:30:54 +09:00
陈大虾🦞
1ffbe09a6a
fix(model): infer provider from allowlist for bare model IDs to prevent prefix drift ( #48369 )
2026-04-04 17:30:54 +09:00
Peter Steinberger
2906cfd6d7
fix: auto-register image-capable config providers ( #51418 ) (thanks @xydt-610)
2026-04-04 17:29:54 +09:00
xydt-610
1d8bba7e39
fix(media-understanding): auto-register image capability for config providers with image input ( #51392 )
2026-04-04 17:29:54 +09:00
Peter Steinberger
3da187156f
docs: clarify native and proxy request shaping
2026-04-04 09:29:09 +01:00
Peter Steinberger
f4855baf35
fix(ci): await async provider test registration
2026-04-04 09:28:43 +01:00
Peter Steinberger
4812b9d2e2
fix: preserve qualified chat model refs ( #49874 ) (thanks @ShionEria)
2026-04-04 17:28:28 +09:00
ShionElia
683c028553
fix: preserve qualified provider prefix in Control UI model selector
...
When sessions report an already-qualified model id (e.g. ollama/qwen3:30b),
resolveServerChatModelValue was re-qualifying it using modelProvider,
producing incorrect values like openai-codex/qwen3:30b.
Preserve already-qualified model refs as-is before applying provider prefix.
Adds test coverage for qualified model preservation.
Fixes #49839
2026-04-04 17:28:28 +09:00
Peter Steinberger
1fcb2cfeb5
docs: clarify provider attribution behavior
2026-04-04 09:27:31 +01:00
Peter Steinberger
73572e04c1
fix: preserve generic DashScope streaming usage ( #52395 ) (thanks @IVY-AI-gif)
2026-04-04 17:25:33 +09:00
Peter Steinberger
a192f345d4
docs: refresh key-free web search ordering
2026-04-04 09:25:20 +01:00
Peter Steinberger
54cfd746de
docs: polish moonshot kimi docs ( #57883 ) (thanks @chenxin-yan)
2026-04-04 17:23:29 +09:00
Chenxin Yan
8347022b50
remove redundency
2026-04-04 17:23:29 +09:00
Chenxin Yan
6615c5788b
docs: fix incorrect Kimi Coding provider ID and model refs
...
The Kimi Coding plugin registers with provider ID `kimi` and default
model ID `kimi-code`, making the correct model ref `kimi/kimi-code`.
The docs incorrectly showed `kimi-coding/k2p5` as the provider/model
ref. This is confusing because `kimi-coding` is only a plugin alias,
not the actual provider ID used in config.
Updated all references in:
- docs/concepts/model-providers.md
- docs/providers/moonshot.md
- docs/zh-CN/concepts/model-providers.md
- docs/zh-CN/providers/moonshot.md
2026-04-04 17:23:29 +09:00
Peter Steinberger
df4b5d2137
docs: refresh self-hosted web search references
2026-04-04 09:22:30 +01:00
Vincent Koc
cdccbf2c1c
fix(github-copilot): send IDE auth headers on runtime requests ( #60755 )
...
* Fix Copilot IDE auth headers
* fix(github-copilot): align tests and changelog
* fix(changelog): scope copilot replacement entry
---------
Co-authored-by: VACInc <3279061+VACInc@users.noreply.github.com >
2026-04-04 17:22:19 +09:00
Peter Steinberger
38ed8c355a
docs: refresh perplexity web search references
2026-04-04 09:21:06 +01:00
Vincent Koc
e4c3df2fb6
docs(changelog): note cache boundary fix
2026-04-04 17:20:24 +09:00
Vincent Koc
a50b838dc2
test(agents): annotate cache trace wrapper params
2026-04-04 17:20:23 +09:00
Vincent Koc
1a13c34f5b
fix(agents): close cache boundary transport gaps
2026-04-04 17:20:23 +09:00
Peter Steinberger
58a56d9a82
feat: add MiniMax TTS provider ( #55921 ) (thanks @duncanita)
2026-04-04 09:19:45 +01:00
Peter Steinberger
a746f0e8c3
style: normalize telegram fetch test formatting
2026-04-04 09:19:45 +01:00
Vincent Koc
7ad43f21d3
style(msteams): format split graph message import
2026-04-04 09:19:45 +01:00
gnuduncan
e934211170
fix(minimax): use global TTS endpoint default and add missing Talk Mode overrides
...
Switch DEFAULT_MINIMAX_TTS_BASE_URL from api.minimaxi.com (CN) to
api.minimax.io (global) so international API keys work out of the box.
Add vol and pitch to resolveTalkOverrides for parity with resolveTalkConfig.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-04-04 09:19:45 +01:00
gnuduncan
7d7f5d85b4
feat(minimax): add native TTS speech provider (T2A v2)
...
Add MiniMax as a fourth TTS provider alongside OpenAI, ElevenLabs, and
Microsoft. Registers a SpeechProviderPlugin in the existing minimax
extension with config resolution, directive parsing, and Talk Mode
support. Hex-encoded audio response from the T2A v2 API is decoded to
MP3.
Closes #52720
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-04-04 09:19:45 +01:00
Peter Steinberger
49d962a82f
docs: refresh brave web search references
2026-04-04 09:19:11 +01:00
Peter Steinberger
d1a4363783
fix(runtime): restore gateway watch on legacy state
2026-04-04 09:18:28 +01:00
Peter Steinberger
0051a86b8f
docs: clarify synthesized web search count behavior
2026-04-04 09:17:49 +01:00
Gaston Rodriguez
b6b1d5dd6c
Moonshot: reuse native base URL for Kimi web search
2026-04-04 17:16:29 +09:00
Peter Steinberger
e5d03f734a
docs: refresh kimi web search setup
2026-04-04 09:15:01 +01:00
Peter Steinberger
21ca006eca
fix(infra): restore approval account binding compatibility
2026-04-04 09:13:11 +01:00
Peter Steinberger
af7c6f4c68
fix: harden kimi web search setup ( #59356 ) (thanks @Innocent-children)
2026-04-04 17:11:47 +09:00
innocent-children
216294765b
fix(kimi): unify runtime model fallback to kimi-k2.5
...
Remove DEFAULT_KIMI_MODEL (moonshot-v1-128k) and align resolveKimiModel
fallback to DEFAULT_KIMI_SEARCH_MODEL (kimi-k2.5). The legacy model
does not support the $web_search builtin_function tool, so env-var-only
users without a configured model would hit the original bug.
2026-04-04 17:11:47 +09:00
innocent-children
111495d3ca
修复kimi web_search错误
2026-04-04 17:11:47 +09:00
Peter Steinberger
11a87b4b7a
docs: clarify plugin facade runtime snapshots
2026-04-04 09:11:25 +01:00
Peter Steinberger
85ade25003
docs: refresh minimax multimodal references
2026-04-04 09:09:06 +01:00
Peter Steinberger
daac149744
fix(ci): honor runtime config snapshots for facades
2026-04-04 09:08:25 +01:00
Peter Steinberger
42f6de16b2
fix: advertise MiniMax M2.7 image input ( #54843 ) (thanks @MerlinMiao88888888)
2026-04-04 17:07:35 +09:00
王淼0668000666
51d998d828
minimax: add image capability to MiniMax-M2.7 model
2026-04-04 17:07:35 +09:00
王淼0668000666
87b41ca693
minimax: add image capability to MiniMax-M2.7 model
2026-04-04 17:07:35 +09:00
Peter Steinberger
ed866020df
docs: refresh task reconciliation references
2026-04-04 09:07:08 +01:00
Peter Steinberger
7d1575b5df
fix: reconcile stale cron and chat-backed tasks ( #60310 ) (thanks @lml2468)
2026-04-04 17:05:57 +09:00
Peter Steinberger
7036e5afbf
fix: honor exec approval security from approvals ( #60310 ) (thanks @lml2468)
2026-04-04 17:05:57 +09:00
Peter Steinberger
8cec7c68b9
fix(ci): restore typecheck on main
2026-04-04 09:05:17 +01:00
Peter Steinberger
da50b492c8
docs: refresh gateway status diagnostics refs
2026-04-04 09:05:08 +01:00
Peter Steinberger
dc2575f6c4
docs: clarify local agent plugin preload
2026-04-04 09:04:11 +01:00
Peter Steinberger
7671f4f1e3
docs: clarify gateway and plugin http auth scopes
2026-04-04 09:01:05 +01:00
Peter Steinberger
bc75968074
perf(cli): trim gateway status startup imports
2026-04-04 08:59:56 +01:00
Peter Steinberger
be15805a84
refactor(runtime): lazy-load control-ui and channel-config surfaces
2026-04-04 08:59:56 +01:00
Peter Steinberger
f9e9d4e357
fix(cli): preload plugins for local agent runs
2026-04-04 08:59:37 +01:00
