Fireworks AI - Fast Inference Platform

How Fireworks AI works

Fireworks AI is the fastest inference platform for generative AI, designed to build and run magical AI applications in seconds. The platform provides serverless access to popular open-source models like DeepSeek, Llama, Qwen, and Mistral with optimized speed, high throughput, and minimal latency. Built for developers who need reliable, blazing-fast AI infrastructure without GPU management complexity.

Recommended Models

We recommend good coding models with competitive pricing and high context windows.

For the most updated information, please visit Fireworks AI’s pricing page.

Model	Pricing (1M tokens)	Context Window
Llama 4 Maverick `recommended`	$0.22/$ 0.88	~131k tokens
Llama 4 Scout	$0.15/$ 0.60	~131k tokens
DeepSeek V3	$0.90	~128k tokens
Qwen3 235B	$0.22/$ 0.88	~131k tokens

Creating API Key

Fireworks AI Account is required to create API Key.

Go directly to Fireworks AI Console to create a new API Key. Or, follow these steps:

Visit app.fireworks.ai and create an account or sign in
Once logged in, navigate to the API Keys page in your account settings
Click “Create API Key” button
Give your key a descriptive name (e.g., ‘Kodus’ or any name you prefer)
Click “Create” to generate the key
Copy the API key immediately and save it somewhere secure - you won’t be able to see it again

New accounts come with $1 in free credits to get started with your projects.

How to use

System Requirements

Recommended Hardware

Sized for the default local sandbox mode (kodus-graph + cross-file context run inside the worker container).

CPU: 2+ cores (4+ recommended for repos above ~100k LOC)
RAM: 8GB+ (16GB recommended when running the local sandbox on large repos)
Storage: 60GB+ free space (Postgres holds the AST graph cache; grows with repo count and PR volume)

Required Software

Docker (latest stable) with the Compose plugin
Domain name or fixed IP if you want to receive Git webhooks from cloud providers (GitHub.com, GitLab.com, etc.)

Required Ports

Default host port mappings — adjust in .env if any conflict.

3000 — Kodus Web App
3001 — API
3332 — Webhooks
5432 — PostgreSQL
27017 — MongoDB
5672, 15672, 15692 — RabbitMQ (AMQP, management UI, metrics)
3101 — MCP Manager (only if API_MCP_SERVER_ENABLED=true)

Services

What ./scripts/install.sh brings up, based on your .env.Core (always on):

api — main backend
worker — code-review jobs
webhooks — Git provider webhook receiver
kodus-web — Next.js frontend
db_kodus_postgres, db_kodus_mongodb, rabbitmq — local infrastructure. Skip with USE_LOCAL_DB=false / USE_LOCAL_RABBITMQ=false to point at managed instances.

Optional:

kodus-mcp-manager — Model Context Protocol broker. Enable with API_MCP_SERVER_ENABLED=true. See MCP Manager.
worker-analytics — Cockpit ingestion (DORA metrics, PR classifier). Self-hosted Enterprise only, not wired by default. See Analytics Worker.

Code review uses an AST graph + cross-file context that runs in a sandbox — local (default, runs inside the worker) or e2b (paid remote sandbox). See Sandbox & AST Graph for the modes, caching behavior, and when to pick each.

Internet access is required if you plan to connect to cloud-based Git services (GitHub, GitLab, Bitbucket) or cloud LLM providers (OpenAI, Anthropic, etc.). For self-hosted Git tools and on-prem LLMs within your network, external internet access is optional.

Domain Name Setup (Optional)

If you're planning to integrate Kodus with cloud-based Git providers (GitHub, GitLab, or Bitbucket), you'll need public-facing URLs for both the Kodus Web App and its API. This allows your server to receive webhooks for proper Code Review functionality and ensures correct application behavior. We recommend setting up two subdomains:

One for the Web Application, e.g., kodus-web.yourdomain.com.
One for the API, e.g., kodus-api.yourdomain.com.

Webhooks are handled by a separate service (port 3332). You can either:

Use a dedicated webhooks subdomain, e.g., kodus-webhooks.yourdomain.com, or
Keep using the API domain and route /github/webhook, /gitlab/webhook, etc. to the webhooks service in your reverse proxy.

Both subdomains should have DNS A records pointing to your server's IP address. Later in this guide, we will configure a reverse proxy (Nginx) to route requests to these subdomains to the correct internal services. This setup is essential for full functionality, including webhooks and authentication.

Note: If you're only connecting to self-hosted Git tools on your network and do not require public access or webhooks, you might be able to use a simpler setup, but this guide focuses on public-facing deployments.

Setup

Clone the installer repository

git clone https://github.com/kodustech/kodus-installer.git
cd kodus-installer

Copy the example environment file

cp .env.example .env

Generate secure keys for the required environment variables

./generate-keys.sh

Edit the environment file

Edit .env with your values using your preferred text editor.

nano .env

See Environment Variables Configuration for detailed instructions.

Run the installer

./scripts/install.sh

Success 🎉

When complete, Kodus Services should be running on your machine. You can verify your installation using the following script:

./scripts/doctor.sh

Access the web interface

Once you access the web interface for the first time, you'll need to:

Create your admin account - This will be the first user with full system access
Configure your Git provider - Connect GitHub, GitLab, or Bitbucket following the on-screen instructions
Select repositories for analysis - Choose which code repositories Kody will review

For detailed steps on the initial configuration process, refer to our Getting Started Guide.

Configure Fireworks AI in Environment File

Edit your .env file and configure the core settings. For LLM Integration, use Fireworks AI in Fixed Mode:

# Core System Settings (update with your domains)
WEB_HOSTNAME_API="kodus-api.yourdomain.com"    
WEB_PORT_API=443                               
NEXTAUTH_URL="https://kodus-web.yourdomain.com"

# Security Keys (generate with openssl commands above)
WEB_NEXTAUTH_SECRET="your-generated-secret"
API_CRYPTO_KEY="your-generated-hex-key"
API_JWT_SECRET="your-generated-secret"
API_JWT_REFRESH_SECRET="your-generated-secret"

# Database Configuration
API_PG_DB_PASSWORD="your-secure-db-password"
API_MG_DB_PASSWORD="your-secure-db-password"

# Fireworks AI Configuration (Fixed Mode) 
API_LLM_PROVIDER_MODEL="accounts/fireworks/models/llama4-maverick-instruct-basic"  # Choose your preferred model
API_OPENAI_FORCE_BASE_URL="https://api.fireworks.ai/inference/v1"                  # Fireworks AI API URL  
API_OPEN_AI_API_KEY="your-fireworks-api-key"                                       # Your Fireworks AI API Key

# Git Provider Webhooks (choose your provider)
API_GITHUB_CODE_MANAGEMENT_WEBHOOK="https://kodus-api.yourdomain.com/github/webhook"
# or API_GITLAB_CODE_MANAGEMENT_WEBHOOK="https://kodus-api.yourdomain.com/gitlab/webhook"
# or GLOBAL_BITBUCKET_CODE_MANAGEMENT_WEBHOOK="https://kodus-api.yourdomain.com/bitbucket/webhook"

Webhook URLs must reach the Webhooks service (port 3332). Use a dedicated webhooks domain or route /.../webhook to port 3332 in your reverse proxy.

Fixed Mode is ideal for Fireworks AI because it provides OpenAI-compatible APIs with blazing-fast inference speeds and access to cutting-edge open-source models with zero setup time.

Run the Installation Script

Looking for more control? Check out our docker-compose file for manual deployment options.

Set the proper permissions for the installation script:

chmod +x scripts/install.sh

Run the script:

./scripts/install.sh

What the Installer Does

Our installer automates several important steps:

Verifies Docker installation
Creates networks for Kodus services
Clones repositories and configures environment files
Runs docker-compose to start all services
Executes database migrations
Seeds initial data

🎉 Success! When complete, the Kodus Web App and backend services (API, worker, webhooks, MCP manager) should be running on your machine. You can verify your installation by visiting http://localhost:3000 - you should see the Kodus Web Application interface.

Code Review features will not work yet unless you complete the reverse proxy setup. Without this configuration, external Git providers cannot send webhooks to your instance.

Set Up Reverse Proxy (For Production)

For webhooks and external access, configure Nginx:

# Web App (port 3000)
server {
    listen 80;
    server_name kodus-web.yourdomain.com;
    location / {
        proxy_pass http://localhost:3000;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}

# API (port 3001)  
server {
    listen 80;
    server_name kodus-api.yourdomain.com;
    location ~ ^/(github|gitlab|bitbucket|azure-repos)/webhook {
        proxy_pass http://localhost:3332;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }

    location / {
        proxy_pass http://localhost:3001;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}

Verify Fireworks AI Integration

In addition to the basic installation verification, confirm that Fireworks AI is working:

# Verify Fireworks AI API connection specifically
docker-compose logs api worker | grep -i fireworks

For detailed information about SSL setup, monitoring, and advanced configurations, see our complete deployment guide.

Troubleshooting

API Key Issues

Verify your API key is correct and active in Fireworks AI Console
Check if you have sufficient credits in your Fireworks AI account
Ensure there are no extra spaces in your .env file
New accounts receive $1 in free credits to get started

Model Not Found

Check if the model name is correctly spelled in your configuration
Verify the model is available in Fireworks AI’s current model library
Try with a different model from our recommended list
Check the Fireworks AI models documentation

Connection Errors

Verify your server has internet access to reach api.fireworks.ai
Check if there are any firewall restrictions
Review the API/worker logs for detailed error messages
Ensure you’re using the correct API endpoint

Performance Issues

Fireworks AI provides industry-leading speeds with minimal latency
Check your network connectivity for optimal performance
Consider using dedicated deployments for enterprise workloads
Monitor your usage patterns to optimize API calls

Rate Limiting

Fireworks AI provides high rate limits on serverless infrastructure
Check your current usage in the Fireworks AI dashboard
Consider upgrading to dedicated deployments for higher throughput
Contact Fireworks AI support for enterprise rate limit adjustments

Overview

Code Review

Business Logic Validation

Rules & Automation

Setup & Integrations

Troubleshooting

LLM Providers

Fireworks AI - Fast Inference Platform

How Fireworks AI works

Recommended Models

Creating API Key

How to use

System Requirements

Domain Name Setup (Optional)

Setup

Configure Fireworks AI in Environment File

Run the Installation Script

What the Installer Does

Set Up Reverse Proxy (For Production)

Verify Fireworks AI Integration

Troubleshooting

​How Fireworks AI works

​Recommended Models

​Creating API Key

​How to use

​System Requirements

​Domain Name Setup (Optional)

​Setup

​Configure Fireworks AI in Environment File

​Run the Installation Script

​What the Installer Does

​Set Up Reverse Proxy (For Production)

​Verify Fireworks AI Integration

​Troubleshooting

How Fireworks AI works

Recommended Models

Creating API Key

How to use

System Requirements

Domain Name Setup (Optional)

Setup

Configure Fireworks AI in Environment File

Run the Installation Script

What the Installer Does

Set Up Reverse Proxy (For Production)

Verify Fireworks AI Integration

Troubleshooting