🧾 Changelog

[0.8.3] - 2025-10-12

Added

  • Accurate cost estimation support using TokenFlow.AI embedded registry data.
  • JSON deserialization shim for model registry compatibility across versions.
  • Comprehensive XML documentation for PromptStreamService.AnalyzePrompt().

Changed

  • CLI packaging updated with PackAsTool=true and ToolCommandName=promptstream.
  • Improved registry loading and cost calculation reliability.

Fixed

  • Zero-cost issue in promptstream analyze command when running as global tool.

Verified

promptstream analyze --template "Explain neural networks simply." --model gpt-4o

Displays accurate estimated cost output.

v0.8.1

  • Added new analyze command for token and cost analysis
  • Updated documentation (cli-analyze.md)
  • Minor CLI help and formatting improvements

v0.8.0 (Upcoming)

  • 🧩 Aligned PromptStream.AI with Flow.AI.Core v0.2.0
  • 🔧 Removed local duplicate models and interfaces
  • ✅ Refactored PromptBuilder and PromptContextManager to use shared DTOs
  • 🧪 Updated all tests for Flow.AI.Core models — 100% coverage restored
  • 📘 Added new documentation page core-integration.md
  • ⚙️ Prepared v0.8.0 NuGet release workflow

v0.7.0

  • Added PromptStream.AI.CLI with build, validate, generate, and context commands.
  • Added full test suite (100% code coverage).
  • Introduced DI-ready TokenFlowModelClient.
  • Polished prompt validation and async handling.

v0.6.x

  • Added PromptStreamService orchestration layer.
  • Introduced PromptValidator and PromptContextManager.