What is an LLM router?
An LLM router acts as a proxy layer that directs requests to various language models. Developers use it to aggregate multiple providers into a single endpoint. This approach simplifies codebases by hiding vendor-specific quirks.
However, routers introduce latency and potential points of failure. They often route traffic based on cost, latency, or model availability. Some routers also apply their own content filters before sending requests to the underlying model.
For many use cases, a dedicated single-model endpoint is more reliable. It eliminates the routing logic and reduces overhead. You get direct access to the model's capabilities without intermediary processing.
This distinction matters for applications requiring low latency or specific model behaviors. Routers are useful for multi-model strategies. Dedicated endpoints are better for focused, high-volume workloads.
Why use an uncensored model?
Uncensored models remove restrictive content filters that often trigger on adult or controversial topics. Standard models may refuse lawful requests due to broad safety policies. Uncensored variants allow creative freedom and accurate roleplay.
These models are ideal for content creation, adult entertainment, and research. They handle nuanced prompts without false refusals. The model does not infer context from commercial branding to decide what is acceptable.
Privacy is another key factor. Many providers train on user data. Our service does not use prompts for training. Your data remains yours. This is crucial for proprietary or sensitive applications.
Content limits still apply. We block sexual content involving minors. Everything else is fair game. This includes mature themes, violence, and political opinions.
Open Router vs. our dedicated router
Open Router aggregates many models. You pay per model per request. Routing logic can add latency. Our service offers a single, dedicated uncensored model. This ensures consistent performance and pricing.
Open Router requires you to manage multiple model IDs. We use a single ID: "uncensored". This simplifies code. You do not need to switch models based on task difficulty.
Pricing on Open Router varies by model. Our pricing is fixed at $0.15 per 1M input tokens and $0.50 per 1M output tokens. This predictability helps with budgeting. You know exactly what each request costs.
Our service is designed for developers who want reliability. We do not route between vendors. We run one optimized model. This reduces complexity and improves response times.
Drop-in compatibility explained
Our API is fully OpenAI-compatible. You can use the same SDKs and libraries. Change the base URL and API key. Your code remains unchanged. This applies to Python, Node.js, and other supported languages.
The endpoint is /v1/chat/completions. It accepts standard parameters like temperature and top_p. Function calling works as expected. JSON mode is supported for structured outputs.
Streaming via SSE is available. Token usage is reported in the final chunk. This matches the OpenAI streaming protocol. You can integrate it into existing pipelines without modification.
This compatibility extends to CLI tools. Tools like Open Code can switch to our endpoint easily. Just update the configuration. No custom adapters are needed.
Cost comparison: Open Router vs. RouterLLM
Open Router charges per model. Prices vary widely. High-end models can cost significantly more. Our pricing is uniform. $0.15 per 1M input tokens. $0.50 per 1M output tokens.
This is often cheaper than using large models on Open Router. It is more expensive than using small models. But you get a high-quality uncensored model. The quality is consistent.
We use prepaid crypto credits. There are no monthly fees. Errors are free. You only pay for successful completions. This protects your budget from wasted requests.
Bonus credits are available for larger top-ups. 5% bonus for $50. 10% for $100. 15% for $200. This reduces the effective cost per token. It rewards high-volume users.
Technical setup: changing the base URL
To switch to our API, update the base URL in your SDK configuration. Point it to HTTPS://api.routerllm.cc/v1. Use your generated API key for authentication.
Set the model ID to "uncensored". This ensures you are using our dedicated model. All other parameters remain standard. Temperature, top_p, and stop sequences work as expected.
Context window is 100,000 tokens. Max output is 16,000 tokens. Default max tokens is 2,048. Adjust these settings based on your needs. Larger outputs require higher limits.
Rate limits are 300 requests per minute. 8 concurrent requests per key. Request body size is limited to 8 MB. These limits are generous for most applications. Monitor your usage to stay within bounds.
Handling streaming and function calls
Streaming is supported via Server-Sent Events (SSE). Each chunk contains partial text. The final chunk includes token usage. This allows real-time updates in your UI.
Function calling works with standard tool definitions. Define tools in your request. The model will output JSON for tool calls. Parse the response to execute functions.
JSON mode ensures structured output. Use response_format: {"type": "json_object"}. The model will return valid JSON. This is useful for data extraction and API integrations.
Parameters like temperature and top_p control randomness. Use seed for reproducibility. Presence and frequency penalties help avoid repetition. Experiment with these settings to find the best balance.
Privacy and data usage
We collect minimal data. An email address is required for account creation. No phone number is needed. You can sign in with Google or use email and password.
Prompts are not used for training. Your data is not sold. This ensures confidentiality for sensitive applications. You retain full ownership of your inputs and outputs.
Payment information is handled via crypto. Transactions are recorded on the blockchain. This adds an extra layer of privacy.
Refunds are not available for credit. Mistakes like double charges are fixed via support. Credit never expires. This provides flexibility for long-term projects.
Getting started with your first request
Sign up for a free account. You receive $0.50 in trial credit. The trial lasts for 7 days.
Generate your API key on the dashboard. Use this key in your SDK configuration. Set the base URL to HTTPS://api.routerllm.cc/v1.
Make a simple request. Send a prompt to the /v1/chat/completions endpoint. The model will return a response. Check the output for accuracy and tone.
Top up with crypto for higher limits. Use USDT (TRC20) or USDC (Base). Enter the amount between $20 and $500. Bonus credits are applied automatically. Start building your uncensored application today.
