How to Use Claude Code with Ivern: Complete Guide

Claude Code is a powerful AI coding assistant, but using it effectively often requires manual task switching between projects. Ivern transforms Claude Code from a tool into a team member — coordinating it with other AI agents for automated workflows.

This guide shows you how to connect Claude Code to Ivern, build squads, and orchestrate AI-powered coding workflows.

Why Connect Claude Code to Ivern?

Claude Code excels at writing code, but it works alone. Ivern lets you:

• Orchestrate Claude Code with other agents — Combine Claude Code with OpenAI, Cursor, or custom agents in coordinated teams.
• Automate multi-stage workflows — Have agents research requirements, write code, run tests, and review results automatically.
• Eliminate manual handoffs — No more copying Claude's output into other tools. Agents pass work between each other seamlessly.
• Watch work in real-time — See Claude Code's decisions and progress as they unfold.
• Scale across your team — Share AI squads with your team for consistent, documented workflows.

Real-World Impact

Before Ivern:

You manually:
1. Open Claude Code
2. Explain the task
3. Copy Claude's code
4. Paste into your IDE
5. Run tests manually
6. Open another AI tool for review
7. Copy review feedback
8. Apply changes manually

With Ivern:

You submit a task once:
1. Ivern squad automatically: Research → Code → Test → Review
2. Watch real-time streaming as agents collaborate
3. Review final output with all stages documented
4. Integrate tested, reviewed code

Prerequisites

Before connecting Claude Code to Ivern, you'll need:

1. Anthropic API Key — Get your key from console.anthropic.com
2. Ivern Account — Sign up free at ivern.ai/signup
3. Claude Code Access — Ensure you have access to Claude Code through Anthropic

API Key Setup

1. Go to Anthropic Console
2. Navigate to "API Keys"
3. Create a new API key
4. Copy the key (you won't see it again)
5. Keep it secure — never share publicly

Step-by-Step: Connecting Claude Code to Ivern

Step 1: Sign Up for Ivern (Free)

1. Navigate to ivern.ai/signup
2. Create your account (email, password)
3. Verify your email
4. Complete onboarding

Time: 2 minutes

Step 2: Connect Your Anthropic API Key

1. From the Ivern dashboard, go to "Settings" → "Agent Connections"
2. Click "Connect New Agent"
3. Select "Claude Code" from the provider list
4. Paste your Anthropic API key
5. Click "Connect"

Ivern validates your key and adds Claude Code to your available agents.

Time: 1 minute

Step 3: Choose a Claude Code Role

Ivern provides pre-built agent role templates optimized for different tasks. For Claude Code, choose from:

• Coder — Optimized for writing, reviewing, and debugging code across multiple languages
• Researcher — Focused on finding best practices, documentation, and similar implementations
• Reviewer — Specialized in code quality checks, security reviews, and testing coverage
• Project Manager — Coordinates coding tasks, breaks down features, and tracks progress

Select the role that matches your intended use case.

Step 4: Create Your First Squad

A "squad" is a team of AI agents that work together on tasks.

1. Go to "Squads" in Ivern
2. Click "Create New Squad"
3. Name your squad (e.g., "Development Squad")
4. Add your Claude Code agent with the selected role
5. Add other agents if desired (OpenAI, Cursor, custom agents)
6. Define the workflow:
   • Sequential: Agents work one after another (e.g., Researcher → Coder → Reviewer)
   • Parallel: Agents work simultaneously (e.g., multiple Coders working on different features)
7. Save your squad

Time: 3 minutes

Step 5: Assign Your First Task

1. Go to your squad's task board
2. Click "New Task"
3. Describe what you want in plain language
4. Assign to the squad
5. Submit

Ivern routes the task through your workflow. Claude Code receives the task at the right stage and begins working.

Time: 2 minutes

Practical Examples: Claude Code Workflows in Ivern

Example 1: Feature Development Workflow

Squad: Feature Development Squad

Agents:

1. Researcher (Claude Code)
2. Coder (Claude Code)
3. Reviewer (Cursor)

Workflow: Sequential

Task: "Add refresh token support to JWT authentication"

Execution:

1. Researcher (Claude Code):
   - Finds JWT best practices
   - Locates similar implementations
   - Gathers security requirements
   
2. Coder (Claude Code):
   - Receives research from Researcher
   - Implements refresh token endpoint
   - Follows project patterns
   - Includes error handling
   
3. Reviewer (Cursor):
   - Reviews implementation
   - Checks for security issues
   - Validates test coverage
   - Provides feedback

You watch all 3 stages happen in real-time streaming.

Example 2: Bug Fix Workflow

Squad: Bug Fix Squad

Agents:

1. Investigator (Claude Code)
2. Coder (Claude Code)
3. Tester (OpenAI)

Workflow: Sequential with validation

Task: "Fix authentication timeout bug on mobile"

Execution:

1. Investigator (Claude Code):
   - Analyzes error logs
   - Reproduces the bug
   - Identifies root cause
   
2. Coder (Claude Code):
   - Receives investigation findings
   - Implements the fix
   - Adds defensive coding to prevent similar issues
   
3. Tester (OpenAI):
   - Tests the fix
   - Runs regression tests
   - Confirms bug is resolved
   - Reports if other issues exist

Each stage passes output to the next. You see the entire workflow unfold.

Example 3: Code Refactoring Workflow

Squad: Refactoring Squad

Agents:

1. Analyzer (Claude Code)
2. Refactorer (Claude Code)
3. Documenter (OpenAI)

Workflow: Sequential

Task: "Refactor user authentication module for better testability"

Execution:

1. Analyzer (Claude Code):
   - Analyzes current module structure
   - Identifies coupling issues
   - Suggests refactoring opportunities
   
2. Refactorer (Claude Code):
   - Receives analysis
   - Performs refactoring
   - Maintains backward compatibility
   - Adds missing tests
   
3. Documenter (OpenAI):
   - Updates API documentation
   - Adds code comments
   - Creates migration guide
   - Records refactoring decisions

Advanced Techniques

Technique 1: Multi-Provider Squads

Mix Claude Code with other AI providers in the same squad:

Example: Research + Code + Review

- Researcher: OpenAI (for web search capabilities)
- Coder: Claude Code (for code generation)
- Reviewer: Cursor (for code quality)

Claude Code does what it does best (write code), while other agents handle research and review.

Technique 2: Parallel Workflows

Have multiple Claude Code agents work simultaneously:

Example: Parallel feature development

- Coder A: Feature X (Claude Code)
- Coder B: Feature Y (Claude Code)
- Coder C: Feature Z (Claude Code)
→ Consolidator: Merges all features

Multiple Claude Code instances work independently, then a consolidator agent merges their work.

Technique 3: Task Chaining

Chain related tasks to maintain context:

Task 1: "Design database schema for user profiles"
Task 2: "Generate TypeScript types based on schema"
Task 3: "Create API endpoints using types"
Task 4: "Build React components consuming API"

Each task uses output from the previous one, maintaining context throughout.

Technique 4: Context Injection

Provide project context to improve Claude Code's output:

When submitting a task:

## Project Context
- Tech stack: Next.js, TypeScript, Prisma
- Architecture: API-first, functional components
- Conventions: camelCase, file-based routing
- Performance requirements: <200ms API response time

## Task
[Your task description]

Claude Code produces better code when it understands your project structure.

Best Practices for Claude Code in Ivern

1. Define Clear Task Boundaries

Vague requests produce vague code.

Bad: "Improve the auth system"

Good: "Add refresh token support to the JWT authentication system. Update /auth/refresh endpoint to accept a refresh token, validate it, and return a new access token. Include error handling for expired tokens. Update TypeScript types accordingly."

2. Provide Relevant Context

Claude Code writes better code with context:

• Tech stack and architecture
• Coding standards and patterns
• Relevant file contents
• Performance requirements
• Dependencies and constraints

3. Use Appropriate Roles

Match the role to the task:

• Coder — For writing, implementing, and refactoring code
• Researcher — For finding best practices, documentation, examples
• Reviewer — For quality checks, security reviews, testing
• Project Manager — For breaking down features and coordinating tasks

4. Validate Before Integrating

Never assume Claude's code is perfect:

• Review generated code manually
• Run tests and linting
• Check for security issues
• Verify functionality matches requirements
• Document edge cases

5. Maintain Context Across Sessions

Use task chaining for related work:

Task A: Design schema
Task B: Generate types (uses Task A output)
Task C: Create endpoints (uses Task B output)
Task D: Build UI (uses Task C output)

Each task builds on the previous one, maintaining context.

Troubleshooting Common Issues

Issue: Claude Code produces incorrect code

Possible causes:

• Insufficient context
• Vague task description
• Missing requirements or constraints

Solutions:

• Add more project context (tech stack, architecture, patterns)
• Be more specific about requirements
• Include examples of desired output
• Break complex tasks into subtasks

Issue: Workflow not executing as expected

Possible causes:

• Incorrect workflow configuration
• Agents not receiving each other's output
• Parallel agents conflicting

Solutions:

• Verify workflow is sequential or parallel as intended
• Check agent roles match workflow stages
• Test with simpler tasks first
• Review real-time streaming to identify bottlenecks

Issue: Integration with existing codebase is difficult

Possible causes:

• Project patterns not documented
• Missing context about coding standards
• Claude Code doesn't understand project structure

Solutions:

• Provide more context in task descriptions
• Add comments explaining project architecture
• Share examples of well-written code
• Use Project Manager agent to coordinate context gathering

Pricing and Costs

Ivern Costs

• Free tier: 15 tasks (including Claude Code tasks)
• Pro tier: $29/month for unlimited tasks and squads
• BYOK: Bring your own Anthropic API key — zero markup

Anthropic API Costs

• Claude 3.5 Sonnet: $3 per million input tokens, $15 per million output tokens
• Claude 3 Opus: $15 per million input tokens, $75 per million output tokens
• Typical coding task: ~10,000-50,000 tokens total
• Typical task cost: $0.03-$0.75 depending on model and complexity

Total cost per task with Ivern: Anthropic API cost only (Ivern free tier) or +$29/month (Ivern Pro) + Anthropic API cost

Cost Savings

Without Ivern:

• Manual task switching: 30-60 minutes per task
• Inconsistent quality
• No audit trail
• Higher API costs from re-prompting

With Ivern:

• Orchestrated workflows: 2-5 minutes setup, automated execution
• Consistent quality with review stages
• Full audit trail documented
• Optimized API usage with agent coordination

Next Steps

After Connecting Claude Code

1. Create more squads — Build specialized squads for different use cases (feature development, bug fixes, refactoring)
2. Experiment with workflows — Try sequential and parallel workflows to find what works best
3. Integrate with your team — Share squads with developers for consistent AI-powered workflows
4. Track performance — Monitor which workflows produce the best results and iterate

Advanced Tutorials

Once you're comfortable with basic Claude Code orchestration:

• Multi-provider squads: Mixing Claude Code, Cursor, and OpenAI
• Advanced workflow patterns: Parallel execution and dynamic routing
• Building custom agents: Extending Ivern with REST API

Summary

Connecting Claude Code to Ivern transforms it from a tool into a team member:

• Setup in 5 minutes — Sign up, connect API key, create squad
• Orchestrate with other agents — Combine Claude Code with OpenAI, Cursor, custom agents
• Automate multi-stage workflows — Research → Code → Test → Review
• Watch real-time collaboration — See decisions and progress unfold
• Scale across teams — Share squads for consistent, documented workflows

Quick Start Checklist

• Sign up at ivern.ai/signup
• Get Anthropic API key from console
• Connect Claude Code to Ivern
• Choose an agent role (Coder, Researcher, Reviewer)
• Create your first squad
• Submit your first task
• Watch real-time streaming
• Review and integrate output

Ready to orchestrate Claude Code? Get started free at ivern.ai/signup.

FAQ

Q: Do I need coding skills to use Claude Code with Ivern?

A: No coding required. Ivern provides a no-code web interface. You define tasks in plain language, and Ivern handles the orchestration.

Q: Can I use multiple Claude Code agents in the same squad?

A: Yes. You can add multiple Claude Code agents with different roles to the same squad. For example, one as Researcher, one as Coder, one as Reviewer.

Q: Does Ivern mark up my Anthropic API costs?

A: No. Ivern uses BYOK (Bring Your Own Key) model. You bring your own Anthropic API key and pay direct provider pricing — zero markup.

Q: Can I mix Claude Code with other AI providers?

A: Yes. Ivern supports cross-provider squads. You can have Claude Code, OpenAI, Cursor, and custom agents all working together.

Q: What happens if Claude Code makes an error?

A: If an agent produces invalid output, the workflow can include validation stages. A Reviewer agent can catch errors and trigger rework. You can also manually review and re-submit tasks.

Q: Is my data secure?

A: Yes. Your API keys are encrypted and stored securely. Ivern never stores your conversations or proprietary data — only task metadata and workflow orchestration.

Start orchestrating Claude Code today at ivern.ai/signup.