@amodalai/amodal 0.1.15 → 0.1.16

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 CHANGED
@@ -1,5 +1,27 @@
1
1
  # @amodalai/amodal
2
2
 
3
+ ## 0.1.16
4
+
5
+ ### Patch Changes
6
+
7
+ - [#69](https://github.com/amodalai/amodal/pull/69) [`407b935`](https://github.com/amodalai/amodal/commit/407b93586178fa19d7c6162f03e259039df336c4) Thanks [@gte620v](https://github.com/gte620v)! - Add prompt caching, multi-model eval comparison, and new provider support
8
+ - Anthropic prompt caching: system prompt and tools sent with cache_control, 90% input cost savings on cache hits
9
+ - Cache-aware cost tracking throughout eval system with savings display
10
+ - Multi-model eval comparison: run evals against multiple models side-by-side with color-graded time/cost table
11
+ - Per-eval history with assertion breakdown, model info, and collapsible UI
12
+ - DeepSeek and Groq provider support via OpenAI-compatible endpoints
13
+ - Configurable eval timeout (20s–300s slider)
14
+ - Tool results now visible in eval output for judge verification
15
+ - Improved judge prompt for specific, evidence-based failure reasoning
16
+ - Auth/rate-limit errors surfaced with actionable UI messaging
17
+ - ConfigWatcher no longer triggers reload spam from eval result writes
18
+ - Session reuse during eval runs to minimize MCP reconnections
19
+
20
+ - Updated dependencies [[`407b935`](https://github.com/amodalai/amodal/commit/407b93586178fa19d7c6162f03e259039df336c4), [`f99b2a1`](https://github.com/amodalai/amodal/commit/f99b2a1d836ee4f57a335182897af696bfce9502)]:
21
+ - @amodalai/core@0.1.16
22
+ - @amodalai/runtime@0.1.16
23
+ - @amodalai/runtime-app@0.1.16
24
+
3
25
  ## 0.1.15
4
26
 
5
27
  ### Patch Changes