Question 1

What is TokenRex?

Accepted Answer

TokenRex is one API for 163+ AI models — including GPT-6 Astra, Claude Opus 5.5, Claude Fable 5.1, Gemini 3.8 Flash and GPT Live 1 — at market price. One key, one wallet, every model.

Question 2

Who is TokenRex for?

Accepted Answer

Developers, startups, students and teams in India who want frontier AI models without a foreign credit card, separate provider accounts or USD billing.

Question 3

What is GPT Live 1?

Accepted Answer

GPT Live 1 is OpenAI's full-duplex voice model: it listens and speaks at the same time, handles interruptions naturally, and delegates reasoning and tool calls to a backend agent. It is audio-only — no image or video input. See the GPT Live 1 model page for the full capability list.

Question 4

How much does GPT Live 1 cost on TokenRex?

Accepted Answer

$0.05 per minute — the official $0.05/minute price — billed per second with no round-up. The backend reasoning model and any tools it calls are billed separately at their own market rates.

Question 5

Which voices does GPT Live 1 support?

Accepted Answer

Twelve named voices — Quartz, Ripple, Vesper, Willow, Stone, Gleam, Meridian, Bossa, Tempo, Beacon, Delta and Cinder — plus the default marin and authorized custom voices. Accents range from Australian and British to Brazilian Portuguese and Southern US.

Question 6

Can GPT Live 1 make phone calls?

Accepted Answer

Yes. It connects over WebRTC, WebSocket or SIP, and official partner integrations include LiveKit, Twilio, Telnyx and Daily/Pipecat, so it can power customer-support and reservation phone agents.

Question 7

What is the difference between Lyria 3 Clip and Lyria 3 Pro?

Accepted Answer

Clip makes fixed 30-second clips, loops and previews ($0.04 per clip). Pro composes full songs up to about 3 minutes with intro, verse, chorus, bridge and outro structure ($0.08 per song). Both match the market price.

Question 8

Can Lyria 3 sing my own lyrics?

Accepted Answer

Yes. Paste your lyrics with section tags like [Verse 1], [Chorus] and [Bridge] in the prompt and Lyria sings them. It can also write its own lyrics, sing in a described vocal style, or produce instrumental-only tracks.

Question 9

How do I control the music Lyria 3 generates?

Accepted Answer

Everything goes in the prompt text: genre, BPM, key/scale, mood, instruments, energy, arrangement and even timing of musical events (for example "drop at 20 seconds"). You can also send an image to set the mood. There are no separate parameter fields.

Question 10

Is TokenRex a model provider?

Accepted Answer

No. TokenRex is an API gateway. Your requests go to the original model providers, and we handle keys, billing and usage tracking for you.

Question 11

Is TokenRex really at market price?

Accepted Answer

Yes. Every model is priced at its market rate. Model pages and your Dashboard show the market price next to what you actually pay.

Question 12

Is TokenRex affiliated with OpenAI, Anthropic or Google?

Accepted Answer

No. TokenRex is an independent service. All model names and trademarks belong to their respective owners.

Question 13

Can I use TokenRex in production?

Accepted Answer

Yes. Use per-key spending limits and model permissions to keep production usage under control, and watch the Status page for live service health.

Question 14

Do I need to be in India?

Accepted Answer

No. The API works worldwide. Pricing in INR and UPI payments are optimised for users in India.

Question 15

Does TokenRex have a status page?

Accepted Answer

Yes. tokenrex.ai/status shows live latency for the TokenRex API, the AI model gateway.

Question 16

How can I contact support?

Accepted Answer

Email hello@tokenrex.ai or use the chat widget on the website.

Question 17

Is there a Playground?

Accepted Answer

Yes. After signing in, the Playground lets you chat, talk with GPT Live, and create images, videos and music with any model before writing code.

Question 18

Can I use TokenRex on mobile?

Accepted Answer

Yes. The website, Dashboard and Playground work on phones, and the API works from any app.

Question 19

Where are the terms and privacy policy?

Accepted Answer

See tokenrex.ai/terms, tokenrex.ai/privacy and tokenrex.ai/refunds.

Question 20

Can I use TokenRex in Claude Code?

Accepted Answer

Yes. Claude models are served in the Anthropic Messages format. Set ANTHROPIC_BASE_URL to https://tokenrex.ai/api/public and ANTHROPIC_AUTH_TOKEN to your TokenRex key, then pick a Claude model such as anthropic/claude-opus-5-5. See the Docs for the full setup.

Question 21

Can I use TokenRex in the Claude desktop app or claude.ai?

Accepted Answer

Partly. The Code section of the Claude desktop app runs Claude Code and can use TokenRex through ~/.claude/settings.json. Regular Claude chats and claude.ai still use your Anthropic account and cannot use a custom API address. The Features page has the complete Code setup.

Question 22

Can I use TokenRex in OpenAI Codex?

Accepted Answer

Yes. Add a custom provider named tokenbaazar in ~/.codex/config.toml with base_url https://tokenrex.ai/api/public/v1 and env_key TOKENBAZAR_API_KEY, then set that environment variable to your key. The Docs page has a copy-ready config.

Question 23

Can I use TokenRex in Apple Xcode?

Accepted Answer

Yes, in Xcode versions that support internet-hosted model providers. Add a provider in Xcode's Intelligence settings with the root URL https://tokenrex.ai/api/public (without /v1), paste your TokenRex key without adding Bearer, then choose a model. The Features page has the full setup.

Question 24

Can I use TokenRex in Cursor?

Accepted Answer

Yes. In Cursor's model settings, override the OpenAI base URL with https://tokenrex.ai/api/public/v1, add your TokenRex key and add the model IDs you want.

Question 25

Can I use TokenRex in VS Code, Cline, Roo Code, Continue or Kilo Code?

Accepted Answer

Yes. Choose the OpenAI-compatible provider, enter https://tokenrex.ai/api/public/v1 as the base URL and your TokenRex key, then type a model ID.

Question 26

Can I use TokenRex in Zed, JetBrains AI, Aider or OpenCode?

Accepted Answer

Yes. Each accepts a custom OpenAI-compatible provider. Use https://tokenrex.ai/api/public/v1 with your TokenRex key and an exact model ID. Open the tool on the Features page for its full setup.

Question 27

Can I use TokenRex with n8n or LangChain?

Accepted Answer

Yes. Use the OpenAI-compatible credential or class, set the base URL to https://tokenrex.ai/api/public/v1 and add your TokenRex key. Copy-ready steps are available on the Features page.

Question 28

Can I use TokenRex in ChatGPT or the Gemini app?

Accepted Answer

No. Consumer apps like ChatGPT and Gemini use their own accounts. TokenRex works with developer tools and your own apps.

Question 29

Which model should I use for coding?

Accepted Answer

GPT-6 Astra, Claude Opus 5.5 and Claude Fable 5.1 are strong choices for hard coding work. Claude Sonnet 5 and Gemini 3.8 Flash are faster and cheaper for everyday tasks. Compare them on the Models page.

Question 30

How do I stop coding agents from overspending?

Accepted Answer

Create a separate key for each tool and set a spending limit and allowed models on that key from the API Keys page.

Question 31

How do I sign up?

Accepted Answer

Click Get started and sign up with email or Google. Verify your email and you can create keys and top up.

Question 32

Can I sign up with Google?

Accepted Answer

Yes. Google sign-in is available.

Question 33

Can I enable two-factor authentication?

Accepted Answer

Yes. Turn on two-factor sign-in from Settings to protect your account.

Question 34

What if I forget my password?

Accepted Answer

Use Forgot password on the sign-in page and we will email you a reset link.

Question 35

Can I change my email or name?

Accepted Answer

Yes, from Settings.

Question 36

Can I delete my account?

Accepted Answer

Yes. Contact support from your account email.

Question 37

Can I have multiple accounts?

Accepted Answer

One person should use one account. Multiple accounts used for abuse are not allowed.

Question 38

Can I invite my team?

Accepted Answer

Accounts are individual today. You can create separate keys for each teammate or project, each with its own limits.

Question 39

Is there a referral program?

Accepted Answer

Yes. Share your referral link from the Referrals page in your account.

Question 40

How do I top up my wallet?

Accepted Answer

Open the Wallet page, enter an amount and pay with Razorpay (UPI, cards, netbanking).

Question 41

What payment methods are accepted?

Accepted Answer

UPI, credit and debit cards and netbanking through Razorpay.

Question 42

Can I set up auto top-up?

Accepted Answer

Yes. UPI Autopay can recharge your wallet automatically when your balance runs low.

Question 43

Can I use coupons?

Accepted Answer

Yes. Enter a coupon code at checkout. Coupons may have minimum amounts, caps, usage limits and expiry dates.

Question 44

How long does a top-up take?

Accepted Answer

Your wallet is credited as soon as the payment is confirmed — usually within seconds.

Question 45

What if money was deducted but my wallet was not updated?

Accepted Answer

Wait a few minutes, then contact support with your payment ID. Payments are credited only once after confirmation.

Question 46

Can I get an invoice?

Accepted Answer

Yes. Every payment has a downloadable receipt on the Transactions page.

Question 47

What currency is my wallet in?

Accepted Answer

INR for users in India. International users see USD.

Question 48

Can I transfer balance to another account?

Accepted Answer

No. Balance cannot be transferred between accounts.

Question 49

What happens when my balance runs out?

Accepted Answer

Requests return a 402 error until you top up or Autopay recharges your wallet.

Question 50

Can I get a refund?

Accepted Answer

Yes, for unused paid wallet balance. Email refunds@tokenrex.ai with your account email and payment ID. When a refund is approved, the amount is deducted from your wallet immediately and reaches your original payment method within up to 3 business days. Spent balance and promotional credit are not refundable. Full details: tokenrex.ai/refunds.

Question 51

Are promotional credits refundable?

Accepted Answer

No. Coupon and promotional credits are not refundable or transferable.

Question 52

What is the base URL?

Accepted Answer

https://tokenrex.ai/api/public/v1 for OpenAI-compatible SDKs. Anthropic SDKs and Claude Code use https://tokenrex.ai/api/public.

Question 53

How do I create an API key?

Accepted Answer

Open API Keys, click Create, name the key and copy it. The full key is shown only once.

Question 54

Can I restrict a key to specific models?

Accepted Answer

Yes. In Manage, allow or block individual models for each key.

Question 55

Can I set a spending limit on a key?

Accepted Answer

Yes. Each key can have its own spending limit.

Question 56

How do I block or delete a key?

Accepted Answer

Open Manage on the API Keys page and block, unblock or delete it. Changes apply immediately.

Question 57

What SDKs can I use?

Accepted Answer

Any OpenAI-compatible SDK — Python, Node, TypeScript, Go, Java, Kotlin, Swift, C#, PHP, Ruby, Rust and more — plus Anthropic SDKs for Claude. The Docs page has examples in 13 languages.

Question 58

Do you support streaming?

Accepted Answer

Yes. Responses stream token by token using server-sent events.

Question 59

Do you support function calling?

Accepted Answer

Yes. Your own function tools are supported on chat models that support tools.

Question 60

Do you support web search or hosted provider tools?

Accepted Answer

Not currently. Hosted provider tools such as web search are not available; use your own function tools instead.

Question 61

Do you support images and PDFs as input?

Accepted Answer

Yes. Vision-capable models accept images, and chat models accept PDF documents.

Question 62

Can I change temperature and other settings?

Accepted Answer

The chat API uses each provider's recommended defaults. Send the model, messages and optionally stream.

Question 63

What HTTP errors can I get?

Accepted Answer

401 invalid key, 402 insufficient balance, 403 model not allowed for the key, 429 rate limited, and 5xx for temporary upstream problems.

Question 64

Can I call the API with cURL or Postman?

Accepted Answer

Yes. Use Authorization: Bearer <your key> with the base URL. The Docs page has cURL examples.

Frequently asked questions