Blog

How a crypto top-up works here: the console quotes an amount, you send it, the balance credits in USD, and every request bills per token from it. What is stored, what is not, and what a dollar buys today.

What no-KYC means for a model API, what we ask for at sign-up and what we store, and why no-KYC and no-log are two different promises.

Why one base URL and one balance beats a console per developer: the models on the key today, how the price compares with each developer’s list, and what switching costs in code (one string).

Input, cached-input and output prices for every Claude model on the key, next to Anthropic's list, and what a chat turn, a long document and a month of each cost. Pay by card or crypto, no KYC.

Input, cached-input and output prices for every GPT model on the key, next to OpenAI's list, with a chat turn, a long document and a month of each costed. Crypto or card, no subscription.

Input, cached-input and output prices for every Gemini model on the key, next to Google's list, and the cost of a chat turn, a long document and a month. Card, Apple Pay, PayPal or crypto.

Input, cached-input and output prices for every Grok model on the key, next to xAI's list, and the cost of a chat turn, a long document and a month of each. No KYC, no subscription.

Input, cached-input and output prices for every DeepSeek model on the key, next to DeepSeek's list, and what a chat turn, a long document and a month cost. Pay by card or crypto.

Input, cached-input and output prices for every Kimi model on the key, next to Moonshot AI's list, and what a chat turn, a long document and a month cost on each.

Input, cached-input and output prices for every GLM model on the key, next to Z.ai's list, and the cost of a chat turn, a long document and a month of each. Crypto or card.

Input, cached-input and output prices for every Qwen model on the key, next to Qwen's list where there is one, and what a chat turn, a long document and a month cost.

Input, cached-input and output prices for every MiniMax model on the key, next to MiniMax's list, and the cost of a chat turn, a long document and a month of each on one prepaid balance.

Input, cached-input and output prices for every MiMo model on the key, next to Xiaomi's list, and what a chat turn, a long document and a month cost. Card or crypto, no KYC.

GPT Image 2 priced per image at every quality and resolution setting on the key, the edit model beside it, and what a month of product shots costs. Paid by card or crypto.

Every FLUX model on the key priced per image: FLUX.2 Pro and Flex, Kontext Pro and Max, FLUX 1.1 Pro and Ultra, each next to Black Forest Labs’ own rate.

Wan video and image models priced per second and per clip, 2.2 through 3.0 Prime, with the cost of a 5, 10 and 15 second clip and the list rate beside each.

Kimi K3.1 Pro on one key: its per-token rate next to Moonshot's list, a chat turn, a long document and a month of agent work costed, and the request that calls it.

Input, cached-input and output prices for every chat model on the key next to the developer's list, a blended price per workload, and the calculator that prices your own token counts.

What OpenRouter charges when you buy credit, what the free models and the bring-your-own-key allowance include, and the same $100 of usage priced on one key here.

What the Claude, ChatGPT, Cursor, Kimi, GLM and Qwen coding plans cost and include, and what the same monthly money buys as tokens on the models they name.

A month of Claude Code on a Claude plan against the same month of turns priced per token on Claude Opus, Sonnet and Haiku here, with the cache reads both sides bill.

Cursor's plans against the same agent work priced per token on a key: one turn late in a session, a steady working day, and the four settings that point Cursor at the key.

OpenCode, Cline, Zoo Code and Kilo Code next to Claude Code: what each tool charges, the same agent turn priced on five models, and one key that runs all five.

An Anthropic key from the Claude Console, or one key that reaches the Claude models and every other developer on this catalog; the steps, the key types, and what a request costs.

Free tiers, trial credit and rate-limited models compared on what they ask for and what they keep, next to a prepaid key from $10 with no subscription and no identity check.

What an API reseller is, which developer terms allow reselling and which do not, three ways to sell on a prepaid key, the margin on one customer's month, and five checks buyers should run.

The usage row one request writes here, the bodies it never writes, and three checks you can run on your own traffic: the signed receipt, the request id and the statement.

The uncensored and abliterated builds on the key, grouped by base model with live per-token prices, what 100 roleplay messages cost on each, and which to pick for which job.

Abliteration removes the direction in a model’s weights that produces a refusal. Where the method comes from, what it changes, what it leaves behind, and the abliterated builds you can call.

How the major chat APIs treat adult fiction, what an uncensored build changes, and how to call one from your app, SillyTavern or Janitor with a key and a base URL.

Which builds suit SillyTavern, Janitor AI and Chub, the settings each frontend sends, and what 100 messages cost on every uncensored model here.

A character card field by field: PNG against JSON, what {{char}} and {{user}} do, how many tokens a card costs, and five cards of ours you can import or call here.

Six alternatives to Character AI compared on account, your own cards, model choice and the price of a reply, including a roleplay chat in the browser on an uncensored build.

Frontends compared on the proxy setting, the account, your own cards and the price of a reply, plus the Janitor fields that keep the frontend and swap only the model.

What OpenRouter's public model list carries under uncensored and roleplay, its per-token prices, and the abliterated builds on one key here with the same kind of pricing.

The uncensored Qwen builds on the key, what each is tuned for, the live price per million tokens, and what 100 roleplay messages cost on each.

What DeepSeek ships in the weights, what the host filters add, and which DeepSeek builds on this key are stock, uncensored or abliterated, with prices for each.

What Heretic does to Gemma’s weights, the Gemma 4 Heretic and roleplay builds on the key, and what a thousand chat turns cost on each.

The uncensored image models on the key, the live price per image at each setting, the safety-checker flag that turns the vendor filter off, and one request that calls them.

The uncensored video models on the key, the live price per second at 480p, 720p and 1080p, and a request that returns the clip with its charge already set.

The security-tuned and uncensored builds on the key ranked by what they answer, with the price of a 40-turn triage session and the model ids to plug into your tools.

The base URL and key settings that point an OpenAI-compatible agent at these models, which id to use for a scan, and what one engagement costs per token.

The open security models that answer by API with per-token prices, next to the hosted plans security teams usually compare, and a request that calls each.