v1.5.1

This release makes Gepetto discoverable by hcli, IDA's plugin manager.

What's Changed

• ida-plugin.json: fix syntax by @williballenthin in #116
• Decoupling tool call handling from tool code by @JusticeRage in #118
• fix(tool): sanitize names and wrap results/errors by @quippy-dev in #121
• feat(ui): add reasoning stream and log categories by @quippy-dev in #117
• fix: address compatibility issues and runtime errors by @quippy-dev in #125
• Thread-safe IDA tooling refresh by @quippy-dev in #126
• feat(ui): introduce resizable splitter for conversation and log views by @quippy-dev in #127
• feat(models): async model discovery & menu refresh by @quippy-dev in #128
• fix(models): set Gemini function_response call id by @quippy-dev in #129
• feat(ida): add new IDA tools and thread-safe helpers by @quippy-dev in #131
• Localize untranslated Gepetto strings across locales by @JusticeRage in #132

v1.5.0

New features

• This release starts the shift towards agentic reverse-engineering by exposing many tools to the LLM, available via the cli.
• A new window dedicated to Gepetto was added (by @quippy-dev, who also vastly improved Gemini support)
• Many minor improvements and bugfixes

What's Changed

• Added support for GPT-4o-mini by @nonetype in #48
• feat: update novita models and readme by @jasonhp in #53
• Add support to LM Studio by @D3fau4 in #58
• Added support for DeepSeek-Chat by @marsharinco in #52
• Fix crash when Ollama or LMStudio isn't configured by @devnoname120 in #61
• Add support for openrouter by @felipejfc in #62
• Add support for Azure OpenAI models by @0xa13d in #66
• fix: the keyword of this parameter should be ‘proxy’ not ‘proxies’ by @jindaxia in #68
• Add three providers by @jindaxia in #69
• Add some features, and some fix by @jindaxia in #71
• Add kluster.ai Support by @themacexpert in #72
• Load available models has better error handling 🚀✨ by @mahmoudimus in #75
• Improve translation installation by @mahmoudimus in #76
• Add deepseek-r1 support by @marsharinco in #80
• Support for Gemini models by @Xplo8E in #83
• UI for renaming variables by @Albeoris in #85
• Add AI-assisted function commenting in Hex-Rays pseudocode by @Albeoris in #86
• Add localization for AI-generated comments by @Albeoris in #87
• Ensure gettext strings present in all locales by @JusticeRage in #89
• Enable CLI tool by @JusticeRage in #91
• Fix OpenAI-compatible plugin configuration by @JusticeRage in #92
• Add auto-rename action and function renaming by @JusticeRage in #93
• Add tool for renaming IDA local variables by @JusticeRage in #94
• Add function rename tool for IDA by @JusticeRage in #95
• Add get_ea tool for resolving symbol addresses by @JusticeRage in #97
• Add list_symbols tool for enumerating IDA symbols by @JusticeRage in #98
• Add decimal to hex conversion tool by @JusticeRage in #99
• Add search tool for text and hex pattern lookups by @JusticeRage in #100
• Add disassembly retrieval tool by @JusticeRage in #101
• Fix plugin loading on IDA < 9 (interactive) & update OpenAI API keys link by @Epieos in #102
• feat(ui): add dockable status panel with streaming output by @quippy-dev in #107
• add metadata for plugin repository by @williballenthin in #105
• feat(models): migrate Gemini to google-genai with tools/stream by @quippy-dev in #112
• fix(ida): stabilize status panel actions by @quippy-dev in #111
• fix(ui): Qt5/Qt6 groundwork; wire request timers by @quippy-dev in #113
• refactor(models): rework Gemini streaming & tools by @quippy-dev in #114

Full Changelog: v1.4.1...v1.5.0

v1.4.1

Bugfix release (#45)

v1.4

• Local models exposed via Ollama are now supported!
  • Ollama will only appear in the menu if the program is running.
• Added a CLI interface in IDA Pro so that users can converse with the selected model directly.
• Important refactoring under the hood so that menus can be generated dynamically based on all available models.
  • An added benefit is that third parties trying to add support for more models should have an easier time.

⚠ A new dependency was added (ollama), make sure you run pip install requirements.txt again after updating! ⚠

v1.3

Added support for a variety of non-OpenAI models

v1.2

v1.2