How does Code By AI work?

Code By AI is a unified platform that orchestrates leading AI models like Claude, Gemini, and Codex. It allows you to leverage the specific strengths of each model—reasoning, speed, or code knowledge—within a single, seamless workflow.

Do I need separate subscriptions for each AI model?

No. A single Code By AI subscription grants you access to all supported AI agents. You don't need to manage multiple accounts or API keys for different providers.

Absolutely. We prioritize enterprise-grade security. Your code is processed securely and is never used to train our models or the third-party models we access.

Can I switch agents mid-project?

Yes! You can switch between agents instantly. For example, you might use Claude for high-level architecture planning and then switch to Gemini for rapid bug fixing, all within the same session.

How does the pricing compare to official APIs?

Our optimized infrastructure allows us to offer token rates from 1/6th of the official API pricing, making advanced AI development significantly more affordable.

What IDEs are supported?

Code By AI integrates seamlessly with popular environments including VS Code and JetBrains IDEs. We also offer a powerful CLI tool for terminal-based workflows.

What does “code by AI” mean?

“Code by AI” means letting an AI coding agent — Claude Code, Codex or Gemini — write, refactor and review code on your behalf. Code By AI is the platform that connects your existing tools to those agents through a single endpoint, from 1/6 of official API prices.

Can I use it as a cheaper Claude Code or Codex setup?

Yes. Point the official CLI at the Code By AI base URL and keep using the same model names. You pay per token from 1/6 of the official list price instead of a monthly subscription.

Code by AI

Agents for serious shipping

Code By AI — one key for Claude, Codex and Gemini

One line of config, same CLI — from 1/6 of the official API price, no subscription.

Launch workspace Explore models & pricing

No credit card required · pay per token

At a glance

Three reasons to switch

Same terminal, one key, lower spend — that is the entire migration.

1API key

One key and one balance cover Claude, Gemini, and Codex — no three sign-ups, no three payment methods.

3mainstream agents

Claude Code, Codex CLI, and Gemini CLI all point their base URL here, so switching models never means switching tools.

1/6of the official price

Billed per 1M tokens at the rate printed on every model card — no subscription, no minimum spend.

Claude Code1/7.5Codex1/7.5Gemini1/10

Works with the agents you already use

One key, one endpoint: keep the terminal you already like.

Claude Code

Anthropic's own terminal agent

Setup guide →

Codex

OpenAI's coding agent for the terminal

Setup guide →

Gemini CLI

Google's terminal agent for Gemini

Setup guide →

OpenCode

Open source, self-hostable terminal agent

Setup guide →

Lightweight, extensible terminal agent

Setup guide →

One key · 4 protocols

Bring the SDK you already use

The same key and the same balance answer four wire formats. No adapter, no gateway rewrite, no second vendor account.

Anthropic Messages

Claude models

POST /api/agent/v1/messages

Point ANTHROPIC_BASE_URL here and Claude Code, the Anthropic SDK and Cline keep their native protocol — streaming, tool calls and token usage unchanged.

Setup guide →

OpenAI Responses

GPT + Claude models

POST /api/agent/v1/responses

The wire_api = "responses" target for Codex CLI. Reasoning items, tool calls and usage frames pass through untouched.

Setup guide →

OpenAI Chat Completions

GPT + Claude models

POST /api/agent/v1/chat/completions

The endpoint most OpenAI clients already speak: OpenCode, Pi, and anything built on the OpenAI SDK.

Setup guide →

Gemini generateContent

Gemini models

POST /api/agent/v1beta/models/{model}:generateContent

Gemini CLI and the Google SDK send :generateContent and :streamGenerateContent here.

Setup guide →

Claude models answer on the Messages endpoint and on both OpenAI-format endpoints — a claude-* model sent to /v1/chat/completions or /v1/responses is forwarded to the Claude channel and billed at Claude prices. GPT and Codex models answer on the OpenAI endpoints, Gemini models on the Gemini endpoint. We forward the request as-is and never translate between protocols, so only a model that belongs to neither family gets a clear 400 instead of a silent substitution.

One-command setup

From key to working agent in one command

Paste one line into a terminal. It asks for your key, writes each tool's own config, backs up every file it touches, and checks the key. Run it again to add tools; --uninstall undoes what it added.

curl -fsSL https://www.codebyai.net/setup.sh | sh -s -- --claude

Sets upClaude CodeCodexGemini CLIOpenCodePiAll options →

Transparent Pricing

AI Model Token Pricing Comparison

Compare official token pricing with Code By AI's significantly lower rates. Our prices start at 1/6th of the official rates, making advanced AI development more accessible.

models supported

13.3%

of official price, on average

−90%

max saving

View all 20 models

Model

Official → Code By AI · per 1M tokens

You save

claude-haiku-4-5-20251001Claude Code

in$1.00$0.13out$5.00$0.67

87% off

claude-sonnet-5Claude CodeDefault

in$2.00$0.27out$10.00$1.33

87% off

claude-sonnet-5-5Claude Code

in$2.00$0.27out$10.00$1.33

87% off

claude-sonnet-4-6Claude Code

in$3.00$0.40out$15.00$2.00

87% off

claude-opus-5-5Claude Code

in$4.00$0.53out$20.00$2.67

87% off

claude-opus-4-6Claude Code

in$5.00$0.67out$25.00$3.33

87% off

claude-opus-4-7Claude Code

in$5.00$0.67out$25.00$3.33

87% off

claude-opus-4-8Claude Code

in$5.00$0.67out$25.00$3.33

87% off

claude-opus-5Claude Code

in$5.00$0.67out$25.00$3.33

87% off

claude-fable-5Claude Code

in$10.00$2.00out$50.00$10.00

80% off

claude-fable-5-1Claude Code

in$10.00$2.00out$50.00$10.00

80% off

gpt-6-solCodex

in$2.00$0.27out$10.00$1.33

87% off

gpt-6.1-solCodex

in$2.00$0.27out$10.00$1.33

87% off

gpt-5.6-terraCodexDefault

in$2.50$0.33out$15.00$2.00

87% off

gpt-5.6-solCodex

in$4.00$0.53out$20.00$2.67

87% off

gpt-6-astraCodex

in$10.00$1.33out$50.00$6.67

87% off

gemini-3.8-flashGeminiDefault

in$0.75$0.08out$3.75$0.38

90% off

gemini-3.7-flashGemini

in$1.00$0.10out$3.00$0.30

90% off

gemini-3.1-pro-previewGemini

in$2.00$0.20out$12.00$1.20

90% off

gemini-4-argonGemini

in$2.00$0.20out$10.00$1.00

90% off

Prices per 1M tokens in USD. Official prices are the vendors' list prices.

See all models and prices Compare plans

Input from $0.08 per 1M tokens

Volume Discount

Top up more, save up to 15%

Automatic tiered discount applied at checkout — the higher the top-up, the bigger the savings.

Top Up Now

−3%
120+ quota
−6%
300+ quota
−10%
600+ quota
BEST VALUE
−15%
1000+ quota

6 quota units = $1 before discounts. The discount reduces the payment, not the quota; payment processing fees may apply.

GPT IMAGE 2.5

AI Image Generation

Create stunning images from text descriptions with GPT Image 2.5. From concept art to product mockups, bring your ideas to life.

Miniature 3D diorama of a cozy developer workspace with a code monitor, plants and a sleeping cat

Generate anything you can imagine

GPT Image 2.5 delivers photorealistic quality with precise text rendering and style control. Available through your existing API key.

Text-to-image and image-to-image generation
Multiple output sizes from 1K to 4K resolution
Seamless integration with your CodeByAI account

What Developers Are Saying

“Switching between agents for different tasks is a game changer. Claude for architecture, Gemini for quick fixes.”

Alice

Frontend Engineer

“The unified context window is incredible. It actually understands how my whole repo fits together.”

Ben

Backend Developer

“I love how it handles the boring stuff like refactoring and tests automatically. Huge time saver.”

Chen

Full-Stack Engineer

“Being able to access all these top-tier models at this price point is unbelievable value.”

Diana

DevOps Engineer

“The 'natural language' commands are spot on. I just ask it to fix a bug and it does it.”

Eric

Mobile Developer

“Professional server maintenance means I never worry about downtime. Rock solid performance.”

Frank

Lead Architect

FAQ

Frequently Asked Questions

Find answers to common questions about Code By AI and get started with AI-powered coding assistance.

Still have questions?

Can't find the answer you're looking for? Our team is here to help.

Contact our support team

Code By AI — one key for Claude, Codex and Gemini

Three reasons to switch

Works with the agents you already use

Bring the SDK you already use

From key to working agent in one command

AI Model Token Pricing Comparison

Top up more, save up to 15%

AI Image Generation

Generate anything you can imagine

What Developers Are Saying

Frequently Asked Questions

Still have questions?

How does Code By AI work?

Do I need separate subscriptions for each AI model?

Is my code secure?

Can I switch agents mid-project?

How does the pricing compare to official APIs?

What IDEs are supported?

What does “code by AI” mean?

Can I use it as a cheaper Claude Code or Codex setup?