Skip to main content
Reach out to our support team with details about your use case, expected volume, and any latency or throughput requirements. We’ll review your needs and can raise your limits accordingly.
Implement exponential backoff in your client code. When you receive a rate limit error (429) or a server error (503), wait before retrying.

How are tokens counted?

Tokens are pieces of text that our models process. A token is roughly 4 characters for English text. Both input and output tokens count toward your usage limits.
Responses include a usage object with token counts, but not a dollar-cost field:
  • prompt_tokens — all input tokens, including cached input.
  • prompt_tokens_details.cached_tokens — the subset of prompt_tokens served from the prefix cache.
  • completion_tokens — all generated tokens, including reasoning tokens.
  • completion_tokens_details.reasoning_tokens — the subset of completion_tokens used for reasoning.
  • total_tokens — prompt_tokens + completion_tokens.
The nested detail counts are already included in their parent totals, so do not add them again. Cached prompt tokens use the cached-input rate; completion tokens, including reasoning tokens, use the output rate.For streaming (stream: true), set stream_options.include_usage: true to receive a final usage chunk before data: [DONE]. Use the rates on Models to calculate cost.
Yes. The full documentation index is published at https://docs.inceptionlabs.ai/llms.txt. Point your LLM, IDE assistant, or agent at that URL to discover every page in the docs. The docs site does not block crawlers or agent traffic.
No. Mercury 2.5, Mercury 2, and Mercury Edit 2 accept text input only. Image generation and image input are not supported. See /get-started/models for supported input formats per model.
Every new account includes a one-time credit of 100 million free tokens. The credit is shared across all models rather than granted per model. When your free tokens run low, add payment information under Billing to continue using the API. For questions about the free credit that aren’t covered here, contact support.
The API does not include a dedicated MCP (Model Context Protocol) server or endpoint. Mercury 2.5 supports OpenAI-compatible tool calling, so any MCP client or agent framework that lets you configure an OpenAI-compatible backend can use it. Set the base URL to https://api.inceptionlabs.ai/v1, provide your API key, and set the model to mercury-2.5. Tool definitions follow the OpenAI function schema shown in the Tool Use guide.
No. Inception’s API covers chat completions (/v1/chat/completions), fill-in-the-middle completions (/v1/fim/completions), and code edit completions (/v1/edit/completions). There is no /v1/embeddings endpoint or embedding model. Pair Mercury with a separate embeddings provider if your workflow needs vector representations.
Fine-tuning is not available through the public API. There is no fine-tuning endpoint, and the Platform dashboard does not offer a self-serve fine-tuning workflow. If you need a custom-tuned Mercury model, contact support with details about your use case, dataset, and deployment requirements.
No. The chat completions request schema does not accept a seed parameter, and temperature on mercury-2 and mercury-2.5 is restricted to the range 0.5–1.0. Values outside that range are silently reset to the model’s default (0.75 for mercury-2, 1 for mercury-2.5) and the response includes a warning field noting the reset. Bit-exact reproducibility across runs is not supported today. Aggregate over multiple samples if you need stable evaluation numbers, or contact support to discuss your use case.The code edit endpoint (/v1/edit/completions) uses a different temperature range of 0–2 with a default of 0.2 and no clamping.
Prefix caching is automatic and requires no request parameter. The usage.prompt_tokens_details.cached_tokens field reports how many prompt tokens were served from cache on a given request, and cached tokens are billed at each model’s input_cache_reads rate (see Models). The API does not expose a documented TTL or a way to pin, invalidate, or inspect cache entries. If your workload depends on a specific retention guarantee, contact support.
Data handling for the Inception API is covered by the Privacy Policy and Terms of Use. Custom data-handling arrangements, including zero-retention configurations for regulated workloads, are handled case by case — contact support with your requirements.