Providers¶

Perspt supports multiple LLM providers through the genai crate.

Supported Providers¶

Provider	Env Variable	Default Model	Notes
OpenAI	`OPENAI_API_KEY`	`gpt-4.1`	GPT-4, GPT-4o, o-series
Anthropic	`ANTHROPIC_API_KEY`	`claude-sonnet-4-20250514`	Claude 4 family
Google Gemini	`GEMINI_API_KEY`	`gemini-3.1-flash-lite-preview`	Gemini 2.x family
Groq	`GROQ_API_KEY`	`llama-3.3-70b-versatile`	Llama, Mixtral
Cohere	`COHERE_API_KEY`	`command-r-plus`	Command R family
xAI	`XAI_API_KEY`	`grok-3-mini-fast`	Grok models
DeepSeek	`DEEPSEEK_API_KEY`	`deepseek-chat`	DeepSeek models
Ollama	(none)	`llama3.2`	Local models via Ollama

1. Environment Variables (recommended):

export GEMINI_API_KEY="your-key"
perspt

2. CLI Flags:

perspt chat --api-key "your-key" --provider-type openai --model gpt-4.1

3. Config File (~/.config/perspt/config.json):

{
  "default_provider": "anthropic",
  "default_model": "claude-sonnet-4-20250514",
  "api_key": "sk-ant-xxx"
}

Priority order: CLI flags > environment variables > config file > auto-detection.

perspt --list-models

OpenAI

export OPENAI_API_KEY="sk-xxx"
perspt chat --model gpt-4.1

Anthropic

export ANTHROPIC_API_KEY="sk-ant-xxx"
perspt chat --model claude-sonnet-4-20250514

Google Gemini

export GEMINI_API_KEY="AIza..."
perspt chat --model gemini-3.1-flash-lite-preview

Ollama (Local)

ollama serve
ollama pull llama3.2
perspt chat --model llama3.2

No API key required. Perspt auto-detects Ollama as the fallback provider.