summaraizer

Summarizes comments from a variety of sources, such as GitHub issues, Slack threads, Reddit threads, and more using AI models from different providers, such as Ollama, Anthropic and OpenAI.

Installation

You can install the CLI by running the following command:

go install github.com/ioki-mobility/summaraizer/cmd/summaraizer@latest

Usage

The usage of the CLI can be found at the project website.

CLI

The minimum required Go version can be found in the go.mod file.

The code is split up into two parts. The source part and the summarization part. First you need to fetch comments from a source, then you can summarize the comments.

Of course, you can also do both independently.

Source aka. fetching comments

The general usage is:

summaraizer [SOURCE] [ARGUMENTS] 

Example sources are:

• github
• gitlab
• slack
• reddit

Summarization aka. calling providers

The general usage is:

summaraizer [PROVIDER] [ARGUMENTS]

Example providers are:

• ollama
• anthropic
• openai

Please note that the provider reads from Stdin as well as require a special JSON format as input:

[
    {
        "author": "Author1",
        "body": "Body1"
    },
    {
        "author": "Author2",
        "body": "Body2"
    }
]

Examples (summarize comments from a source)

Each provided source already respects the JSON format required by the summarization providers. Therefore, you can use piping to chain the commands together.

summaraizer github --issue golang/go/66960 | summaraizer ollama --model llama3

This command fetches the comments of this GitHub issue and summarizes them using the llama3 model via Ollama.

summaraizer reddit --post '/r/ArtificialInteligence/comments/1d16cxl/miss_ai_worlds_first_beauty_contest_with_computer/' | summaraizer openai --model gpt-4o --token SUPER_SECRE_API_TOKEN

This command fetches the comments of this Reddit post and summarizes them using the gpt-4o model via OpenAI.