← $0.50 free · Text + Image + Video · No card

ChatGPT. Midjourney. Runway.
One key replaces them all.

Text, image, and video generation through a single API key and balance. Built for indie builders shipping AI products — not for managing four separate subscriptions.

No subscription. No card. Refunded on failure.

Powering production traffic from

🦅 Cline Cursor 🤖 Claude Code 🦜 Aider 🚀 OpenCode Continue 🔧 Roo Code

The math indie builders already know

Stop paying for tools
you use twice a week.

ChatGPT Plus, Midjourney, and Runway add up to roughly $65/month combined — whether you generate 5,000 messages or five. KodaAPI charges per request. No idle subscriptions.

  • One balance covers chat, image generation, and video generation
  • Unused balance never expires — no "use it or lose it"
  • No monthly minimum, no card required to start
Monthly cost
ChatGPT Plus
Flat fee whether you use it or not
$20/mo
Midjourney
Separate account, separate bill
$30/mo
Runway
Yet another subscription to remember to cancel
$15/mo
KodaAPI
Text + image + video, pay per request
from $0
The same $20 on KodaAPI covers thousands of chat replies, dozens of images, or a handful of videos — your call, not a fixed bundle.

Drop-in compatible

Uses OpenAI format. Change your base URL and API key — your existing code works instantly. No SDK to install, no migration needed.

from openai import OpenAI

client = OpenAI(
  base_url="https://kodaapi.com/v1",
  api_key="YOUR_KODAAPI_KEY"
)

response = client.chat.completions.create(
  model="claude-3-5-sonnet-20241022",
  messages=[{"role": "user", "content": "Hello!"}]
)

print(response.choices[0].message.content)
curl https://kodaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KODAAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      { "role": "user", "content": "Hello!" }
    ]
  }'
# ~/.cursor/config.json · Aider · any OpenAI-compatible agent
{
  "openai_base_url": "https://kodaapi.com/v1",
  "openai_api_key":  "YOUR_KODAAPI_KEY",
  "model":           "gemini-3.5-flash"
}

One Key

Every model worth using.
Text, image, and video.
Same balance.

Chat models, image generators, and video models — switch with a string change, no second invoice. Aliases like best and fast resolve to the right model as your code needs what you mean.

  • OpenAI-compatible /v1/chat/completions, /v1/images/generations
  • Aliases like best and fast auto-resolve to the right model
  • Zero re-tooling — change one line, access every provider
ModelProviderContextIn / Out
gpt-4o OpenAI 128k $5 / $15
claude-3-5-sonnet Anthropic 200k $3 / $15
gemini-3.5-flash Google 1M $1.50 / $9
gpt-4o-mini OpenAI 128k $0.15 / $0.60
gemini-3.1-flash-lite Google 1M $0.25 / $1.50
gemini-2.5-flash Google 1M $0.30 / $2.50
o3-mini OpenAI 128k $1.1 / $4.4
deepseek-v4-pro DeepSeek 1M $0.435 / $0.87
View all 50+ models →

The boring stuff. Handles.

Numbers we can defend
with logs.

50+
models on one key.
Text, reasoning, vision, image & video gen.
8+
active providers built-in.
OpenAI, Anthropic, Google, DeepSeek & more.
0%
user-facing errors.
Last 7 days.
$0.50
free to start.
No card to get started. Cancel anytime.
Live status

Pay only when it works

Refund on failure.
Always.

Holds at submit, charged on success. If the upstream times out, fails, or returns garbage, your balance is restored. Zero engineering required on your side.

Ledger
How a request settles
1
Submit
Hold placed at quoted price.
− hold
2
Settle
Upstream returned tokens. Charged on usage.
− actual
3
Reverse
Upstream timed out or returned an error.
hold released
Every charge has a paired settle or reverse row. No silent retries you can't see.

"1,300 requests a day, 60% cache hit, $20 spent, $24 saved on cached input. The cron runs itself."

T
Thien
Indie Hacker · Vietnam

"One app, every API I needed. Don't waste your life signing up for ten dashboards."

M
Marcus
Full-stack Engineer · Production

No subscription. No expiry.

Top up once.
Use until empty.

Credits never expire. Unused balance is fully refunded if a request fails upstream.

$5
50,000 pts
  • ~5,000 fast-model requests
  • All 20+ models included
  • OpenAI-compatible API
  • Usage dashboard
Get started
$20
210,000 pts +5% bonus
  • ~21,000 fast-model requests
  • All 20+ models included
  • Streaming (SSE) support
  • Full usage analytics
Top up
$100
1,200,000 pts +20% bonus
  • ~120,000 fast-model requests
  • All 20+ models included
  • Best price per token
  • Priority support
Top up

1 pt = 1 token  ·  Credits never expire  ·  Refunded on failure

Be generous with experiments.
Stingy with hours.

The cheap thing in 2026 is the idea that ships first.