Skip to content

CHANGELOG.md

Latest commit

 

History

History
125 lines (71 loc) · 3.43 KB

CHANGELOG.md

File metadata and controls

125 lines (71 loc) · 3.43 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.

[Unreleased]

[0.2.12] - 2025-11-15

Added

  • Tool now supports an post action, which can be used to execute some after llm response is generated. There are following actions supported:
    • output: output the content to stdout (default), let action be empty or output
    • copy: copy the content to clipboard, let action to be copy
    • save: save the content to a file, let action be any file name
    • execute:
      • when action is execute, run, or exec, the content will be treated as a shell command and executed directly.
      • when action is a non-empty string other than above, it will be treated as a shell command template, and the content will be passed as argument to the command. For example, if action is echo, and content is hello, then the command executed will be echo "hello".
    • see
      • tr.toml : a translation tool that copy the translation result to clipboard. gpt -t tr "Hello, how are you?" will copy 你好，你好吗？ to clipboard
      • pa.toml : a command assistant that build shell command from user input and execute it directly. gpt -t pa "list all files in current directory" will execute ls -a command directly.
  • -t or --tool option to specify a tool file to use. -T is changed to set temperature.

[0.2.11] - 2025-10-31

Added

  • Add '--tool' '-t' option to enable a tool use mode.

Tool is a pre-defined system prompt, model, and other configurations to do specific tasks. see Tool for more details.

A example tool 'tr' is located at tr.toml, which translate between chinese and english, you can use it like:

gpt -t samples/tools/tr.toml "Hello, how are you?"

This will output the translation result, 你好，你好吗？

the samples/tools/tr.toml can be copied to $HOME/.gpt/tools/tr.toml, then you can use it like:

gpt -t tr "Hello, how are you?"

[0.2.10] - 2025-10-26

Added

  • Added reasonEffort configuration to control how llm generate response

[0.2.9] - 2024-10-15

Changed

  • Upgraded the openai SDK to v3.
  • Improved changelog format.

[0.2.8]

Changed

  • Verbose level 2 now outputs reason content, and verbose level 3 outputs the raw chunk response.

[0.2.7]

Added

  • Allow using a ProxyMCPClient to expose any HTTP service as an MCP server, for example gpt -M samples/qqwry.mcp.yaml "where is 120.197.169.198's location".

[0.2.6]

Changed

  • Upgraded the openai SDK to v2.

Added

  • Allow referencing MCP prompts by name, for example gpt -M "mcp_url" -s p1 user_prompt.

[0.2.5]

Changed

  • Upgraded the openai SDK to v1.

Added

  • Added streaming HTTP transport support for MCP, indicated by URLs without sse.

[0.2.4]

Added

  • Enabled multi-round MCP calls.

[0.2.3]

Changed

  • Improved compatibility.

Added

  • Support specifying only the model name.

[0.2.2]

Changed

  • Improved compatibility with MCP servers.

[0.2.1]

Fixed

  • Fixed MCP SSE so it starts before use.

[0.2.0]

Added

  • Added support for the Model Content Protocol (MCP); gpt now accepts the -M option to specify the MCP server.