OpenAI-compatible API
Seoffix API docs
POST /api/v1/chat/completions returns one JSON completion. Anonymous callers get 20 requests an hour. A key from pricing raises that to 1,000 a day on Pro or 2,500 a day on Apex.
Need a production key?
Pro is ₹299/month for 1,000 calls a day. Apex is ₹999/year for 2,500 calls a day. The key looks like sfx_live_ plus 48 hex characters.
Code samples
OpenAPI 3.1Claude Desktop, Cursor, and Windsurf
The @seoffix/mcp server exposes DNS, WHOIS, PDF, humanizer, and logo tools. Stdio is for desktop apps. The remote endpoint is https://seoffix.com/api/mcp.
{
"mcpServers": {
"seoffix": {
"command": "npx",
"args": ["-y", "@seoffix/mcp"],
"env": { "SEOFFIX_API_KEY": "sfx_live_your_key_here" }
}
}
} Get an API key Authentication, Rate Limits & Integration Reference
Engineered by Seoffix Tech Lab, the gateway implements the standard OpenAI Chat Completions wire contract at POST /api/v1/chat/completions. Existing Python, Node.js, PHP, Go, and cURL clients connect by overriding only the base URL and passing your sfx_live_ bearer token inside the Authorization HTTP header. Store production keys in server-side environment variables rather than browser bundles so credentials cannot be extracted from client traffic.
HTTP Status Codes & Error Handling
- 200 OK: Valid JSON completion containing
id,model,choices[0].message.content, and token usage counters. - 400 Bad Request: Missing
messagesarray, unsupported model identifier, or malformed UTF-8 request body. - 401 Unauthorized: Bearer key is malformed, expired, or does not match an active Pro Studio or Apex Annual subscription.
- 429 Too Many Requests: Hourly anonymous quota (20 calls/hr) or daily key quota (1,000 Pro / 2,500 Apex) reached; back off until window reset.
- 502 Bad Gateway: Upstream inference cluster timeout; retry idempotently with exponential jitter starting at 1,000 ms.
MCP Transport Modes & Model Routing
For local IDE agents in Cursor, Claude Desktop, and Windsurf, run npx -y @seoffix/mcp over stdio so DNS, RDAP, SSL, PDF, and content tools execute straight from your editor context. Hosted cloud agents and serverless webhooks can target https://seoffix.com/api/mcp over TLS 1.3. Choose task-tuned models such as seoffix-apexnlp for structured extraction, seoffix-codesynth for code generation, and seoffix-toolradar for technical SEO diagnostics.
When do daily API quotas reset?
Authenticated Pro (1,000/day) and Apex (2,500/day) counters reset automatically at 00:00 UTC without manual intervention or overage charges.
Are prompts retained for model training?
No. Gateway payloads are processed transiently to return the JSON completion and enforce quota counters; customer payloads are never used for training.
Where is the machine-readable schema?
Download the OpenAPI 3.1 specification from /docs/openapi to generate typed clients in Postman, Insomnia, or OpenAPI Generator.