AI gateway for agents

Connect Claude Code, Codex, Hermes, and OpenClaw through one API gateway. Keep keys, usage windows, and request costs organized in one place.

Free 14-day trial - no card required. Pick a plan later.

quickstart.ts
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.phantom.si/v1", apiKey: process.env.PHANTOM_API_KEY!, }); const result = await client.chat.completions.create({ model: "your-enabled-model", messages: [{ role: "user", content: "Hello!" }], });
EXAMPLE RESPONSE Example only
{ "choices": [{ "message": { "content": "Hello!" } }] }
sampleChat CompletionsOpenAI-compatible example

One gateway, familiar tools

Keep your existing coding workflow and route compatible requests through Phantom.

Claude CodeCoding agent
CodexCoding agent
HermesAgent framework
OpenClawAgent platform

The gateway at a glance

A quick look at the signals that matter when routing AI workloads.

Usage
1.2MRequests/min
Latency
812 msAvg latency
Caching
92%Cache hit rate
Flexible
15Models served

Illustrative UI values shown for layout preview; replace with live metrics from your backend before publishing.

Built for agents that never stop running.

One set of credentials, scoped keys, and a spend model that keeps surprises away.

One key, every model

A scoped key for each tool. Rotate one without breaking the rest.

Rolling budgets

5-hour bursts and weekly windows with an honest ledger behind every call.

Audited by default

Every request logged with tokens, latency, and exact cost. No guesswork on bills.

Three steps to connect

Create a key, point your agent, and watch the ledger fill in.

01

Create a key

Open the API Keys page and mint a scoped key per tool.

sk_live_...
02

Point your agent

Swap baseURL to https://api.phantom.si/v1 and drop in the key.

baseURL = ...
03

Watch the ledger

Every request lands with tokens, latency, and exact cost.

POST /v1/chat/completions
Simple plans, clear limits

Choose the room you need.

Pick a subscription level, then track your short-window and weekly usage separately. The values below mirror the current landing-page plan data.

Trial access

Trial

Explore the gateway

$0for 14 days
Start free

Usage windows

5-hour burst$7
Weekly window$49
  • 14-day trial
  • Explore compatible tools
  • Usage ledger
5 API keys

Pro

For personal workflows

$6/ month
Choose Pro

Usage windows

5-hour burst$10
Weekly window$70
  • Up to 5 API keys
  • 5-hour burst window
  • Weekly usage window
More room to grow
10 API keys

Scale

For heavier daily use

$10/ month
Choose Scale

Usage windows

5-hour burst$20
Weekly window$140
  • Up to 10 API keys
  • Higher burst allowance
  • Usage and cost visibility
50 API keys

Power

For multi-agent setups

$50/ month
Choose Power

Usage windows

5-hour burst$100
Weekly window$700
  • Up to 50 API keys
  • Largest listed allowances
  • Built for multiple workflows

Subscription prices and usage windows are shown from the current page's plan data, not fetched live from billing. Confirm these values against your pricing/checkout configuration before launch. Model-provider charges, included allowance, and subscription fee should be explained separately in the full pricing page.

Need the full comparison or INR prices?

Compare all 8 tiers

Ready to stop juggling providers?

One key. One ledger. One bill.