Guides

These guides walk you through common workflows and advanced use cases in ProxifAI.

Setting Up a Development Workflow

A complete setup for a team using ProxifAI for the full development lifecycle:

Create an Organization

Set up your organization, invite team members, and configure roles. See Creating an Organization.

Create Projects and Repositories

Create projects for planning and repositories for code. Link them together so issues connect to PRs automatically.

Set Up CI/CD Pipelines

Define YAML-based pipelines in your repositories. Configure secrets, approval gates, and deploy environments.

Configure AI Agents

Set up AI models under Settings → AI Configuration. Configure agent images and compute resources.

Build Automation Workflows

Create workflows in the Flows section to automate repetitive tasks — issue triage, code review, testing, and more.

Common Workflows

Issue-to-PR Flow

Create an issue describing the work
Create a branch referencing the issue
Push code and open a pull request
CI pipeline runs automatically
Team reviews and approves the PR
Merge closes the linked issue

AI-Powered Code Review

Open a pull request
The AI agent is triggered via a workflow
Agent clones the repo, reviews the diff, and leaves comments
Developer addresses feedback
Agent re-reviews and approves

Automated Issue Triage

New issues arrive via the API, CLI, or web
A workflow trigger fires on issue creation
AI agent analyzes the issue content
Agent assigns priority, labels, and a suggested assignee
Team reviews and adjusts as needed

Sprint Planning

Create a new sprint with start and end dates
Review the backlog and drag issues into the sprint
Assign issues to team members
Track progress throughout the sprint
Close the sprint and review completion stats

Integrations

Slack Integration

Connect Slack to receive notifications and interact with ProxifAI:

Receive alerts for PR reviews, pipeline failures, and workflow approvals
Use slash commands to create issues and check status
Interactive buttons for quick actions

Webhook Integration

Set up outbound webhooks for custom integrations:

Subscribe to specific event types
View delivery logs and retry failed deliveries
Test webhooks before going live

AI Gateway as API

Use the AI Gateway as a drop-in replacement for OpenAI-compatible APIs. The gateway is mounted at /api/v1/llm/* on your deployment host:

curl -X POST https://your-host/api/v1/llm/v1/chat/completions \
  -H "Authorization: Bearer pfai_your_token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4.6",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

The AI Gateway ships with three provider types out of the box — openai, anthropic, gemini — plus a generic openai-compatible driver for any OpenAI-shaped endpoint (Ollama, vLLM, OpenRouter, your own proxy). See AI Gateway for the full default model catalog.