Skip to content

What's new with Parakeet

🦜 Parakeet v0.2.9 🥧 [pie]

  • Fixed: function calling: no need to use the JSON format with the query options
  • Added:
  • MCP Streamable HTTP Transport support. MCP HTTP client example: 95-mcp-http
  • squawk.SchemaJSON(schemaJSONString string) *Squawk

🦜 Parakeet v0.2.8 🍩 [doughnut]

  • Landing of Squawk: a Parakeet DSL 
     1
 2
 3
 4
 5
 6
 7
 8
 9
10
    squawk.New().
    Model(model).
    BaseURL(ollamaBaseUrl).
    Provider(provider.Ollama).
    Options(options).
    System("You are a useful AI agent, you are a Star Trek expert.").
    User("Who is James T Kirk?").
    Chat(func(answer llm.Answer, self *squawk.Squawk, err error) {
        fmt.Println(answer.Message.Content)
    })
  • Improvement of the history messages management
  • Added support for structured output to the Docker Model Runner Chat API
  • Added support for structured output to the OpenAI Chat API

🦜 Parakeet v0.2.7 🐳 [spouting whale]

Addition of Docker Model Runner support (and OpenAI at the same time) allowing easy development of generative AI applications in Docker containers.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
modelRunnerURL := "http://model-runner.docker.internal/engines/llama.cpp/v1/"
model := "ai/qwen2.5:latest" 

systemContent := `You are an expert in Star Trek.`
userContent := `Who is Jean-Luc Picard?`

options := llm.SetOptions(map[string]interface{}{
    option.Temperature: 0.5,
    option.RepeatPenalty: 2.0,
})

query := llm.Query{
    Model: model,
    Messages: []llm.Message{
        {Role: "system", Content: systemContent},
        {Role: "user", Content: userContent},
    },
    Options: options,
}

answer, err := completion.Chat(
    modelRunnerURL, 
    query, 
    provider.DockerModelRunner,
)
if err != nil {
    log.Fatal("🫢 Oops!", err)
}
fmt.Println(answer.Message.Content)

Note

👋 Look at this Docker Compose sample examples/82-web-chat-bot-model-runner