Features

Endpoints

OpenAI: Chat Completions, Responses, Embeddings, Images, Audio, Rerank
Anthropic Messages and native Gemini
Conversion between the three protocols in any direction, preserving tool calls, images, thinking and cache usage; streaming and non-streaming
Claude Code and official SDKs work out of the box

Upstream types: OpenAI-compatible (including other relays), Azure OpenAI, Anthropic, Gemini, AWS Bedrock, Vertex AI
Priority + weight routing with automatic failover; when an upstream says "no such model", the request moves to another channel instead of failing
Automatic key suspension and cooldown, scheduled latency probes, long answers never cut off
Outbound address checks against SSRF

Anchored to the official price: site, group and negotiated prices (see Pricing and margin)
Real cost per channel; loss-making prices can't be saved; pricing health check and profit dashboard
Per-token, per-call and long-context tiers; cache reads and writes priced separately
Two-phase ledger: reserve → settle, crash recovery, manual review for unclear cases
Scheduled official price sync and calibration, full price history

Gateway overview: revenue, cost, margin, daily trend, top models and users, channel health
Negotiated pricing: find a user, set "N% of official" per model or in bulk, see the margin live
Billing group decided by API key: dedicated lines and special rates attach to a specific key
Usage logs, settlement review, balance ledger
Menu- and button-level permissions with four preset roles (gateway ops, finance, support, read-only); admin 2FA and IP allowlist
Prometheus metrics

Sign-up and login: username + password, SMS code, email password reset
Self-service API keys: budget, expiry, model scope, IP allowlist; a key is shown only once at creation and only its hash is stored
Model and price list, usage and billing, integration guide

Manual top-up by admins, redemption codes
Online payment: Alipay and aggregated payment providers, extensible; payment secrets are write-only
Referral rewards