Skip to content

CHANGELOG.md

Latest commit

 

History

History
144 lines (108 loc) · 6.6 KB

CHANGELOG.md

File metadata and controls

144 lines (108 loc) · 6.6 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[3.3.0] - 2026-01-15

  • Improved memory usage with incremental file-level updates and memory-efficient contribution caching (#99) - @Sewer56
  • Refactored CLAUDE.md to AGENTS.md with modular skill documentation (#94) - @Sewer56
  • Added support for GPT-5.2-Codex model (#101) - @mike1858

[3.2.2] - 2025-12-27

  • TUI: Display cached tokens in session view instead of redundant tool column (#88) - @Sewer56
  • TUI: Filter summary totals to selected day in session view (#87) - @Sewer56
  • TUI: Add 'r' hotkey to toggle reverse sort order (#84) - @Sewer56
  • Fix hash collisions by using timestamp+id hash for deduplication (#90) - @mike1858
  • Fix file watcher to detect new sessions in nested directories (#86) - @Sewer56
  • Fix OpenCode parsing crash when messages contain boolean summary field (#83) - @Sewer56

[3.2.1] - 2025-12-13

  • Added update notifications: Splitrail now checks GitHub Releases on startup and displays a banner when a new version is available - @mike1858

[3.2.0] - 2025-12-13

  • We now fully support Piebald! Track your Piebald usage across all your favorite providers - @mike1858
  • Breaking: Removed the disk-based caching system entirely to fix stats fluctuation bugs during file watching - @mike1858
  • Added in-memory message cache for fast incremental updates during file watching - @mike1858
  • Removed cache CLI commands (cache stats, cache clear, cache rebuild, cache path) - @mike1858
  • Cold startup is now ~1s instead of instant, but results are guaranteed correct - @mike1858
  • Added timezone support for uploads - @mike1858
  • Fixed Codex CLI duplicate entries by including entry count in globalHash - @mike1858
  • Added GPT-5.2, GPT-5.2-Pro, and GPT-5-Pro model pricing support - @mike1858

[3.1.1] - 2025-12-05

  • Use real GPT-5.1-Codex-Max's pricing (reported pricings should be a bit lower) - @mike1858

[3.1.0] - 2025-12-05

  • We now fully support the Pi coding agent! Track your historical (and current) Pi usage - @mike1858
  • Splitrail Cloud: Update Next.js version in response to critical security vulnerability - @mike1858
  • Splitrail Cloud: support Pi agent - @mike1858

[3.0.0] - 2025-11-30

  • Introducing the VS Code extension: now you can view track token counts, costs and more straight from VS Code - @mike1858
  • Introducing the MCP server: let Claude access your usage history with the Splitrail MCP server - @mike1858
  • Introducing JSON output mode: --json lets you build apps on top of Splitrail - @mike1858
  • 10x the performance: Splitrail startup time has been reduced from ~2s to ~200ms - @mike1858
  • We now fully support OpenCode! Track your historical OpenCode usage - @mike1858
  • Claude Opus 4.5: Anthropic's most powerful frontier model is now in Splitrail - @mike1858
  • Cache manipulation: you can now manipulate Splitrail's cache: view, clear and rebuild.
  • Faster hashing: we've significantly improved hashing speed by moving from SHA256 to xxhash.
  • Fixed an issue where Gemini CLI info messages were incorrectly being parsed as conversation entries - @mike1858
  • Added support for re-uploading zero-cost messages in case we didn't add a model but you used it anyway - @mike1858
  • Added over 3.5 thousand lines of new tests, covering the TUI, upload system, analyzers, utilities, and file watcher - @mike1858

[2.2.3] - 2025-11-23

  • Add code-signing for Windows and macOS builds (#47) - @signadou
  • Fix: Timezone issue in TUI & Add: Gemini CLI reasoning tokens (#45) - @mike1858
  • Feat: Add --full and --force-analyzer flag to upload command (#46) - @mike1858

[2.2.2] - 2025-11-21

  • Correct extraneous upload requests (#43) - @mike1858

[2.2.1] - 2025-11-20

  • Fix REQUEST_ENTITY_TOO_LARGE uploading errors (#41) - @mike1858

[2.2.0] - 2025-11-20

  • Add support for estimated pricing; add GPT-5.1-Codex-Mini and (estimated) GPT-5.1-Codex-Max (#39) - @mike1858
  • Add support for searching across days (#37) - @mike1858
  • Allow content to be an array of content blocks or a string in CC analyzer (#38) - @mike1858
  • Greatly improve performance (#36) - @mike1858
  • Various fixes and improvments to Splitrail Cloud - @mike1858

[2.1.0] - 2025-11-19

  • Add support for Copilot - @mcowger
  • Add better support for Roo Code - @mcowger
  • Add support for per-session token + cost tracking - @mike1858
  • Various fixes and improvements to Splitrail Cloud - @mike1858

[2.0.0] - 2025-11-10

  • Add support for Cline, Kilo Code, Roo Code (Cline forks) and Qwen Code (Gemini CLI fork) (#25 and #26) - @bl-ue
  • Various fixes and improvements to Splitrail Cloud

[1.2.0] - 2025-10-27

  • Correct Claude Code token aggregation for split messages (#23) - @bl-ue

[1.1.3] - 2025-10-24

  • Add support for Claude Haiku 4.5 to the Claude Code analyzer (#21) - @bl-ue

[1.1.2] - 2025-10-01

  • Add support for Claude Sonnet 4.5 to the Claude Code analyzer
  • Ignore the "file-history-snapshot" entry added in Claude Code 2.0.0

[1.1.1] - 2025-09-18

  • Fix Codex CLI (#16) - @bl-ue

[1.1.0] - 2025-09-18

Added

  • Enhanced Codex CLI support with updated analyzer functionality
  • Better formatting and lint compliance

[1.0.1] - 2024-09-14

Added

  • Download link in README for binary releases

Changed

  • Rearranged README structure for better readability
  • Improved documentation and project presentation

Fixed

  • Various formatting and lint error corrections

[1.0.0] - 2025-08-09

Added

  • Initial stable release of Splitrail
  • Real-time automatic uploading to Splitrail Cloud
  • Comprehensive multi-tool support:
    • Claude Code
    • Gemini CLI
    • Codex CLI
  • Rich Terminal User Interface (TUI) using ratatui
  • Advanced cost calculation and token usage analytics
  • File operation tracking (read/write/edit operations with byte/line counts)
  • Comprehensive model support:
    • Claude models (Sonnet 4, Opus 4, Opus 4.1)
    • GPT models (GPT-5 series, GPT-4 series)
    • Gemini models (2.5-pro, 2.5-flash, legacy 1.5 series)
    • Codex CLI models (o3, o1 series, gpt-4.1 series)
  • Configuration management with TOML-based config
  • Splitrail Cloud integration with API token authentication
  • Deduplication logic to prevent duplicate entries
  • Real-time file watching capabilities
  • Parallel processing for performance