@stackbone/cli 0.9.0 → 0.9.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +37 -0
- package/main.js +343 -84
- package/migrations/0122_model_routing_object.sql +16 -0
- package/package.json +1 -1
- package/stackbone-cli-0.9.2.tgz +0 -0
- package/voice-worker/main.js +215 -58
- package/voice-worker/main.js.map +4 -4
- package/stackbone-cli-0.9.0.tgz +0 -0
package/CHANGELOG.md
CHANGED
|
@@ -7,6 +7,43 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
|
|
|
7
7
|
|
|
8
8
|
## [Unreleased]
|
|
9
9
|
|
|
10
|
+
## [0.9.2] - 2026-11-10
|
|
11
|
+
|
|
12
|
+
### Added
|
|
13
|
+
|
|
14
|
+
- **A voice call starts its knowledge search before the caller finishes.**
|
|
15
|
+
With Flux, the voice worker, the one `stackbone dev` runs too, sends the
|
|
16
|
+
speech model's early guess to the box, and the box searches the agent's RAG
|
|
17
|
+
collection on it. A turn whose sentence matches the guess takes the passages
|
|
18
|
+
already found instead of searching again. Nothing is written to the
|
|
19
|
+
conversation and no chat model runs before the turn is real. An agent with a
|
|
20
|
+
model-backed input guardrail gets no early search. Each voice run records
|
|
21
|
+
`voiceKnowledgePrefetch` (`hit`, `miss` or `none`) in its metadata, and each
|
|
22
|
+
turn's metrics carry `preflightLeadMs`, how early the guess left (STA3-94).
|
|
23
|
+
- **The `turn.eagerThreshold` voice setting** (0.3 to 0.7, default 0.3) tunes
|
|
24
|
+
how early that guess comes. Deploy the worker on LiveKit Cloud before the
|
|
25
|
+
boxes: an older worker refuses a call whose settings carry it.
|
|
26
|
+
- **Model routing takes OpenRouter's whole `provider` object** on deep agents
|
|
27
|
+
and subagents, and the voice warm-up pins the same provider. The box applies
|
|
28
|
+
platform migration 0122 on boot.
|
|
29
|
+
|
|
30
|
+
## [0.9.1] - 2026-11-10
|
|
31
|
+
|
|
32
|
+
### Changed
|
|
33
|
+
|
|
34
|
+
- **The thinking sound waits for a long reply.** The voice worker, the one
|
|
35
|
+
`stackbone dev` runs too, starts the thinking sound at once for a box tool
|
|
36
|
+
call, and for any other wait only once it passes the
|
|
37
|
+
`backgroundAudio.thinkingDelayMs` voice setting (default 2000 ms). Before,
|
|
38
|
+
it played on every reply, since every reply waits for the knowledge search
|
|
39
|
+
and the model's first token. Studio's Voice tab gets a **Thinking delay
|
|
40
|
+
(ms)** field. Deploy the worker on LiveKit Cloud before the boxes: an older
|
|
41
|
+
worker refuses a call whose settings carry the delay.
|
|
42
|
+
- **The box's missing-credentials messages name the Voice provider screen.**
|
|
43
|
+
A call or a voice status check on a box with no readable LiveKit
|
|
44
|
+
credentials now sends you to Studio's **Voice provider** screen, where the
|
|
45
|
+
credentials moved (STA3-89).
|
|
46
|
+
|
|
10
47
|
## [0.9.0] - 2026-11-10
|
|
11
48
|
|
|
12
49
|
### Added
|