Skip to content
Agents tracked: 258 Downloads (7d): 219M up 6.1% GitHub stars: 5.5M VS Code installs: 148M Releases (7d): 327 Agent status: 1 with issues Updated Oct 7, 2026

BettaFish changelog: what's new each month

Every stable BettaFish release summarised by month: the highlights, new features, improvements, fixes and anything you need to act on. 3 months covered; the current month updates daily.

December 2025

2 releases: 2.1.0 → 3.0.0

December brought significant expansions to BettaFish's capabilities, including new search and clustering features, while the month culminated in the launch of MiroFish, a companion prediction engine that completes the data-to-decision workflow.

Highlights

  • Insight Agent now clusters and samples search results using semantic analysis to return representative results from each group.
  • Added support for Ansi AI search as a more economical search option.
  • Integrated xhshow library to improve signature generation and request parameters.
  • Released MiroFish, a prediction engine that extends BettaFish's data analysis into forecasting capabilities.
  • Added secure system shutdown functionality.

New

  • DeepSearchAgent.clusterandsampleresults() method for intelligent search result clustering and sampling
  • Ansi AI search integration
  • Safe shutdown system function
  • MiroFish prediction engine release

Improved

  • Signature generation and request parameters through xhshow library integration

November 2025

3 releases: 1.1.0 → 2.0.0

November brought major stability improvements and significant architectural enhancements, culminating in version 2.0.0 which adds full HTML/PDF rendering capabilities and offline dependency support.

Highlights

  • New HTML and PDF rendering pipeline enables command-line report generation without external dependencies.
  • LLM interface management now supports multi-model dynamic switching with unified configuration.
  • Docker deployment fully enhanced with ARM architecture support and automated database setup.
  • Memory usage optimized to under 1GB peak consumption, reducing hardware requirements.
  • Enhanced Agent temporal perception and improved stability across multiple subsystems.

New

  • HTML and PDF renderers with decoupled content and presentation layers.
  • PDF export API endpoint with automatic Pango dependency checking.
  • LLM interface independent configuration supporting multiple models and services.
  • Chart.js validator with LLM-powered automatic repair to prevent frontend issues.
  • Offline dependency integration including fonts, MathJax, and Chart.js.
  • Visualization log console with virtual list and SSE history compensation.
  • Command-line report generation mode without web interface.
  • Dynamic frontend rendering for system parameters and analysis options in Web UI.

Improved

  • LLM call timeout increased to 1800 seconds to reduce timeout errors.
  • Agent temporal perception improved by adding real-world time to LLM prompts.
  • Installation process switched to uv for faster setup and dependency resolution.
  • Dynamic host and port configuration with Pydantic-based validation system.
  • Docker image optimized with improved Dockerfile build parameters.
  • MediaCrawler and ForumEngine modules updated for better communication stability.
  • Health check logic redesigned to resolve TCP connection issues.
  • Forum Engine Agent restructured to improve message extraction and semantic understanding.

Fixed

  • Fixed fetchnews inability to properly retrieve news data.
  • Corrected agent temporal perception error in LLM interactions.
  • Fixed torch and transformers installation issues.
  • Resolved sourcekeyword table missing from bilibilivideo database.
  • Fixed data structure simplification and encoding issues in crawlers.
  • Corrected docker-compose.yml file path problems to prevent configuration loss.

September 2025

1 release: 1.0.0

September brought the initial release of BettaFish with a comprehensive suite of specialized AI agents and real-time content monitoring capabilities. The platform launched with multiple agent types designed for different analysis and research tasks.

Highlights

  • Five specialized agent types: Insight Agent for database analysis, Media Agent for multimodal content, Query Agent for web search, Report Agent for automated report generation, and Forum Engine for collaborative agent interactions.
  • AI crawler cluster operating continuously to capture content across 10+ major social media platforms.
  • Real-time hot topic detection and deep comment analysis across social networks.

New

  • Insight Agent for private database mining and deep analysis.
  • Media Agent with multimodal content analysis capabilities.
  • Query Agent with domestic and international web search functionality.
  • Report Agent with built-in templates for multi-turn report generation.
  • Forum Engine implementing collaborative agent mechanism with chain-of-thought and debate features.
  • AI crawler cluster for 24/7 continuous content collection.
  • Support for Weibo, Xiaohongshu, TikTok, Kuaishou, and 6+ additional social media platforms.
  • Real-time hot content capture with user comment analysis.

Summaries are written automatically from the official release notes (full changelog ↗); check the original notes before relying on a detail. BettaFish: pricing, features and alternatives · All changelogs