CHANGELOG.md

Changelog

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

[0.14.0]

Improved

• Prompting (reduce token usage)

[0.13.0]

Added

• Support for GPT-5 and OpenAI OSS models
• Searchable model selection in configuration UI
• API error messages now displayed to users

Fixed

• Fixed load screen rendering issues
• Fixed custom base_url handler
• Fixed double slash in URL paths

Updated

• Dependency updates:
  • fastapi: 0.116.1 → 0.118.0
  • pydantic-settings: 2.10.1 → 2.11.0
  • uvicorn[standard]: 0.35.0 → 0.37.0
  • openai: 1.101.0 → 2.1.0
  • pydantic: 2.11.7 → 2.11.10
  • cryptography: 45.0.6 → 46.0.2
  • pytest: 8.4.1 → 8.4.2
• GitHub Actions updates (github-script, stale, setup-python)

[0.12.0]

Added

• optional Socks support

[0.11.1]

Fixed

• Removed reasoing effort

Updated

• Requirements

[0.11.0]

Fixed

• Fixed catalog size ; Thanks to @EnriqueSantos-dev #15; fixes #14
• reasoning effort; fixes #16

Removed

• Adult Flag (legacy routes are still available)

Updated

• Requirements

[0.10.0]

Added

• Language selector for TMDB Data; Thanks to @EnriqueSantos-dev #11
• Caching dependent on Language

Refactor

• Config decryption through Depends -> allow access of configured fields earlier in request cycle

[0.9.3]

Cleanup

• Configuration Page Cleanup thanks to @0xConstant1 #8

Fixed

• Model Validation for gpt-5 models
• load correct catalog selection on reconfigure

[0.9.2]

Fixed

• Fixed loading default model env @0xConstant1 #8

Updated

• Requirements

Changed

• Configure/Preview: Add disclaimer that changing settings alters the manifest URL and requires re-adding the addon in Stremio
• Make max results a slider to prevent average users from inputting too high numbers and running into timeouts

[0.9.1]

Changed

• Configure/Preview: Add disclaimer that changing settings alters the manifest URL and requires re-adding the addon in Stremio
• Make max results a slider to prevent average users from inputting too high numbers and running into timeouts

[0.9.0]

Added

• Catalog Selection. Makes it possible to select a catalog to use.

Removed

• Adult search for TMDB. This relaxes the cache a bit.

[0.8.3]

Changes

• Updated Prompt

[0.8.2]

Fix

• hasScheduledVideos is bool

[0.8.1]

Added

• Logos

Fix

• imdbRating as str

[0.8.0]

Added

• Poster URLs are done on a per config base (TMDB or PosterDB)

Fix

• Use IMDB_ID where possible to better work with default stremio clients

[0.7.5]

Added

• slugify for search cache key

[0.7.5]

Added

• Configuration validation on save: tests LLM, TMDB, and RPDB connections
• Meaningful error messages for configuration validation failures

Fixed

• Configuration save now validates all services before creating manifest URLs

[0.7.4]

Fixed

• removed behavior which led to an infinite loop

[0.7.3]

Added

• validation for TMDB key

Fixed

• fix empty cache served for searches

[0.7.2]

Fixed

• fix CACHE catalog:critics_picks

[0.7.1]

Fixed

• get correct CATALOG_PROMPTS

[0.7.0]

Changed

• Cache searches for 2h

[0.6.0]

Changed

• CI/CD complete refactor
• Better Prompting

[0.5.3]

Changed

• CI/CD complete refactor
• Better Prompting

[0.5.3]

Changed

• Configure: Auto-fetch models on first focus with spinner and keep placeholder selected until user chooses

[0.5.2]

Changed

• Configure: Provider first, dynamic API key label per provider, custom uses "LLM API Key" and no help text
• Configure: Remove default model unless provided by existing configuration

[0.5.0]

Added

• Cache for static content
• simplify pagination for catalogs

[0.5.0]

Easier initial setup with builtin support for Openrouter, Gemini, OpenAI and Anthropic models

Added

• Gzip compression for responses
• Browser caching for static content (30-day cache with immutable flag)

Changed

• Refactor Configure page for easier support of multiple model providers
• Refactor Prompt handling (inspired by stremio-ai-search)
• Nicer error message on wrong API key
• General refactoring and duplicate removal

[0.3.6]

Changed

• Minor patches and maintenance.

[0.3.5]

Changed

• Reduce log verbosity: demote noisy INFO logs to DEBUG in API and services; keep critical warnings/errors. Improves ops signal-to-noise.

[0.3.4]

Fixed

• Routes for splitted manifest

[0.3.3]

Changed

• Configure page: support adult flag via URL and redirect endpoint

[0.3.2]

Added

• Dynamic caching

[0.3.1]

Fix

• pagination for Omni

[0.3.0]

Added

• Catalogs caching with configurable limits

Changed

• Cache refactor with stable, config-independent keys
• Mask sensitive config values in logs
• Load config with new URL handling
• Update preview and index templates for cached catalogs
• Dependency updates and fixes

Fixed

• Test stability and minor issues

[0.2.3]

Changed

• Show response time in preview
• Minor refactoring

[0.2.2]

Added

• Background image and footer in UI

Changed

• GitHub Actions workflow updates
• Default LLM model set to horizon-beta
• Improved manifest splitting handling and config extraction
• Better error handling on decryption failures

[0.2.1]

• Github Action push…

[0.2.0]

Added

• Split manifest support for separate movie and series addons
  • Query parameter support: ?type=movie or ?type=series
  • Dedicated endpoints: /manifest/movie.json and /manifest/series.json
  • Web UI dropdown to select addon type (Combined, Movies Only, Series Only)
  • New SPLIT_MANIFESTS environment variable for future configuration
• Comprehensive test coverage for all manifest URL variations
• Backward compatibility maintained for existing manifest URLs

Changed

• Enhanced preview page with manifest type selection dropdown
• Improved manifest builder with configurable content types
• Updated API routes to support multiple manifest formats

[0.1.1]

Added

• Catalog View
  • Show 50 Recommendations from LLM
  • New env MAX_CATALOG_RESULTS

Fixed

• Install to app

[0.1.0]

Added

• TV show support with comprehensive integration
• Worker configuration system with documentation
• Request scheme handling from headers
• New test files for scheme handling
• TMDb service for TV show data retrieval
• Utility functions for data conversion

Changed

• Updated Docker configuration for better worker support
• Improved Stremio API implementation
• Enhanced preview template with TV show display
• Updated README with latest information
• Fixed Docker Compose configuration

Removed

• Debug endpoint from web API

[0.0.4] - 2025-08-01

Changed

• Use .env file for configuration (marked as minor change)

[0.0.3]

Changed

• Updated README (marked as patch)

[0.0.2]

Added

• First public version

[0.0.1]

Added

• Initial project setup