Frequently Asked Questions

Getting started is quick and simple. Create a free account — you'll receive 1,000 free credits immediately. Then generate an API key from your dashboard under "API Keys." Include this key in every request as a Bearer token in the Authorization header. See our API documentation for code examples in Python, JavaScript, and cURL.

Yes — our API is OpenAI-compatible, so migration is minimal. Change the endpoint URL from https://api.openai.com/v1 to ours, and swap your OpenAI key for your LLM Resayil API key. You'll typically save 70–90% on costs. Try it risk-free with your 1,000 free credits — no credit card required.

LLM Resayil targets 99.5% uptime for our API infrastructure. We run redundant servers with automatic failover to handle disruptions. Enterprise customers can purchase dedicated SLA agreements with guaranteed uptime and priority support. Contact our team for custom SLA options.

LLM Resayil uses Bearer token authentication. Include your API key in the Authorization header of every request: Authorization: Bearer YOUR_API_KEY. This is identical to OpenAI's pattern. Keys are secret — never share them publicly or commit them to source control. Create multiple keys for different environments and revoke them any time.

Rate limits are tier-based. Free: 10 req/min. Basic: 100 req/min. Pro: 500 req/min. Enterprise: custom. Exceeding your limit returns HTTP 429 (Too Many Requests). Monitor usage in your dashboard, upgrade your tier, or contact us for an enterprise plan with custom thresholds.

Yes, we fully support streaming via Server-Sent Events (SSE). Instead of waiting for the full response, tokens are sent as they're generated — providing dramatically faster perceived response times. To enable streaming, set stream: true in your request parameters. The format is identical to OpenAI's streaming, so existing code works with minimal changes. See our documentation for examples.

The API returns standard HTTP status codes: 2xx = success, 4xx = client errors, 5xx = server errors. Common errors: 401 (invalid API key), 429 (rate limit exceeded), 400 (bad parameters). Implement proper handling: use exponential backoff for 429s, validate parameters before sending, log error bodies (they contain descriptive messages). See our documentation for the full error code reference.

• Write concise prompts — fewer words means fewer tokens and faster responses
• Use the smallest model that fits your task to save cost and time
• Cache responses for repeated queries in your application layer
• Set max_tokens appropriately to stop generation early
• Enable streaming for better perceived performance in interactive apps
• Monitor token usage in the dashboard and adjust your strategy accordingly

LLM Resayil uses a pay-per-token credit system — no monthly subscriptions. Purchase credit packs starting from 5.00 USD for 500 credits. Tokens are deducted in real-time as you make API calls. Input and output tokens are counted separately at different rates per model. Credits never expire and top-ups are instant. Use the cost calculator to estimate spending.

Your dashboard provides real-time spending and usage monitoring. The Dashboard page shows current credit balance, daily/monthly usage charts, and cost breakdown by model. The Billing section displays transaction history and credit purchases. API responses include token counts so you can track spending in your own logs. Everything updates in real-time as you make calls.

Yes — subscription tier rate limits naturally cap overall consumption. Additionally, implement safeguards in your code: cache responses when possible, use smaller models for simple tasks, and monitor usage continuously. The dashboard shows cost projections to help estimate spending in advance. Contact us for enterprise plans with custom spending caps.

LLM Resayil offers access to 45+ powerful AI models, including Meta Llama, Mistral, NeuralChat, Orca, and more. Each has different capabilities, speeds, and pricing. Smaller models (7B parameters) are faster and cheaper — ideal for simple tasks and high-volume applications. Larger models (70B+) deliver superior reasoning for complex tasks. See our Features page for the full list with benchmarks.

Currently, LLM Resayil provides access to 45+ pre-trained AI models. Custom model hosting and fine-tuning services are available through our enterprise offering — contact our team to discuss dedicated options. Many users find that careful prompt engineering with the right base model solves 90% of use cases without fine-tuning.

A 401 error means your API key is missing, invalid, or expired. Verify the header format: Authorization: Bearer YOUR_API_KEY. Check that you copied the full key — even one missing character causes failure. If the key looks correct, generate a new one from "API Keys" in your dashboard. Also confirm you're using the correct API endpoint URL. See our documentation for authentication details.

Response time depends on model size and request complexity. If you experience slowness:

• Try a smaller, faster model if accuracy permits
• Enable streaming mode for faster perceived response times
• Optimize prompts to be more concise
• Check if you're near your rate limit — throttled requests are slower
• If issues persist, contact support with timing details and request IDs