Ollama Provider Added

Release Notes — v0.12.0

New Features

Ollama Provider Added

• Agentify Toolkit now supports Ollama models via both:
• Ollama Cloud API (API key)
• Local Ollama Server

This continues the theme of flexible and rapid agent experimentation with interchangeable model providers.
Unified Agent Execution
Ollama-backed agents can now be:

# Run interactively:
agentify run examples/agents/ollama.yaml

# Served as an HTTP API:
agentify serve examples/agents/ollama.yaml

# Deployed to the local agent runtime:
agentify runtime start
agentify deploy examples/agents

Environment & Config Improvements

API keys can now be provided through a .env file
.env.example is now part of the repo for quick onboarding

Local development requires only:
cp .env.example .env

then add keys as needed.

Providers now auto-read configuration via python-dotenv without requiring:
agentify provider add <provider>

This command is now deprecated as redundant.

Developer Experience

Provider abstraction is more consistent across OpenAI, Mistral, DeepSeek, Anthropic, XAI, Google, AWS Bedrock, and now Ollama.

Local-first workflows are more streamlined for rapid prototyping and testing.

Deprecations
agentify provider add is now deprecated and will be phased out in a future release.