How do I start using LingCoreX.AI?
Register an account, verify your email if required, sign in to the dashboard, and create an API key. From there you can call the OpenAI-compatible endpoint with your chosen model and your LingCoreX.AI key.
Getting started
The fastest path from sign-up to the first successful API call.
Do I need a special SDK or a custom client?
Usually no. Most teams can keep their existing OpenAI-compatible SDK or tooling and only change the base URL plus the API key. That is the default integration path described in the docs.
Does LingCoreX.AI also support Anthropic-native model groups for Claude Code and similar tools?
Yes. In addition to OpenAI-compatible access, LingCoreX.AI also exposes Anthropic-native request paths for the model groups configured for that interface. This is useful for Claude Code and other Anthropic-compatible tools because they can keep their native request behavior and often deliver a smoother coding workflow than forcing everything through an OpenAI-compatible shim.
Anthropic-native access is intended for tools that expect `/v1/messages` style behavior and provider-specific headers.
Claude Code can be pointed at LingCoreX.AI by setting the Anthropic base URL, API key, and the model option assigned to your key.
Where should I look if I want working examples quickly?
Start with the public docs page. It includes cURL examples, Node.js and Python SDK usage, and notes for common tools that accept OpenAI-compatible providers.
Usage and routing rules
Questions about model selection, request flow, and operational behavior.
How do I choose which model to call?
Choose from the models exposed in your dashboard or current public model list. The model name you send in the request must match an enabled model identifier available to your account or project.
What is a model group?
A model group is the access and billing unit exposed to your API key or team. Instead of granting every key direct access to every underlying model, LingCoreX.AI can bind a key or a team policy to one model group, and that group determines which models are reachable and how that traffic is governed.
Groups help separate use cases such as general chat, coding, premium reasoning, or Anthropic-native tool access.
They are also useful for team-level budgeting, access control, and cleaner key management.
Does LingCoreX.AI send traffic to upstream model providers?
Yes. LingCoreX.AI operates as an aggregation and routing layer. Your request may be relayed to an eligible upstream provider in order to fulfill the selected model call.
Are there rate limits or quota controls?
Yes. Platform-side controls such as request pacing, quota policies, or balance checks may apply depending on account state, team settings, and model configuration.
High-volume integrations should test against expected concurrency before launch.
If a workflow is business critical, monitor responses and retries rather than assuming unlimited throughput.
Billing and metering
How credits, token accounting, and usage visibility work.
How does billing work?
LingCoreX.AI uses a credits-based service model. Requests may go through a pre-consumption check, but the final charge depends on measured token usage and the pricing rules attached to the model and token type.
Why can dashboard usage and billing values differ for a short period of time?
Usage panels can update earlier than final billing aggregation. Near-real-time charts are useful for operational visibility, while final billing may reflect delayed settlement or provider-side accounting updates.
Which token categories matter for billing?
Depending on the model and upstream provider, charges may distinguish between input tokens, output tokens, cache reads, and cache writes. The platform pricing rules determine how those categories are converted into credits.
Cache read tokens usually mean the provider reused an existing prompt cache entry and charged the lower cache-read rate for that reused portion.
Cache write tokens usually mean the provider created or refreshed a cacheable prompt segment and charged the cache-write rate for building that cache state.
Privacy and security boundaries
What the platform can help with and where customer responsibility remains.
Should I send passwords, secrets, or personal identifiers to the model?
As a default operating rule, no. Sensitive data should be minimized before it reaches any third-party model boundary. Planned controls such as privacy masking are intended to reduce accidental disclosure, but customers should still define their own prohibited-data policy.
Who is responsible for request content and compliance decisions?
Customers remain responsible for the data they choose to submit, their lawful basis for processing, and the policies that govern which content may be sent to external model providers. LingCoreX.AI can add operational controls, but cannot decide your regulatory posture for you.
Can teams share usage visibility and budget controls?
Yes. Team-oriented usage controls and centralized visibility are part of the platform direction, so organizations can monitor spend and activity without distributing raw provider credentials to every user.
Support and policy expectations
Operational expectations before you put real traffic on the platform.
Where do I find the platform rules and legal documents?
Read the public terms and privacy pages alongside the docs. Those pages define the service boundaries, data-handling expectations, and the general operating model for the platform.
What should I do if I hit issues during integration?
Start by checking the docs and model list, then verify your API key, endpoint, and model identifier. If the problem is operational rather than configuration-related, contact platform support through the available support channel.