ShadowGate API Docs

A privacy-first inference gateway. One API key, multiple models, no data retention.

Overview

ShadowGate exposes a single inference endpoint that proxies requests to multiple LLM providers (cloud and local). You authenticate with one API key. We log metadata only and never persist your prompts or responses.

Base URL

https://api.thetechcave.net/v1

Authentication

All requests require an API key, sent as a bearer token:

Authorization: Bearer <SHADOWGATE_API_KEY>

Generate and manage keys from your dashboard. Keep your keys secret — treat them like passwords. Revoke any key you believe is compromised.

Inference

POST /v1/inference

Send a chat completion request through the gateway.

curl https://api.thetechcave.net/v1/inference \
  -H "Authorization: Bearer $SHADOWGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "What is the capital of France?"}
    ]
  }'

Request body

• model (string, required) — the model identifier to route to.
• messages (array, required) — the chat messages.
• temperature (number, optional) — sampling temperature.
• max_tokens (integer, optional) — max tokens to generate.
• stream (boolean, optional) — set true for streaming responses.

Response

A standard chat-completion-shaped response containing the generated message and usage metadata (token counts). We do not store the conversation content.

Models

The exact provider list depends on your plan. Pro and Enterprise plans include access to local models in addition to cloud providers. Get the current list with:

curl https://api.thetechcave.net/v1/models \
  -H "Authorization: Bearer $SHADOWGATE_API_KEY"

Usage & billing

Usage is counted by requests and tokens per your plan. Your dashboard shows current usage. Overages beyond your monthly allowance are billed at the published usage rate. You can set a hard monthly cap.

Privacy guarantees

• No content retention. Prompts and responses are never stored.
• Metadata only. We log timestamps, model and usage counts for billing and operations.
• No profiling. Your activity is never used to build a profile of you.

See the Privacy Policy for full details.

Error handling

• 401 Unauthorized — missing or invalid API key.
• 402 Payment Required — billing issue or hard cap reached.
• 429 Too Many Requests — rate limit or plan quota exhausted.
• 5xx — upstream provider or gateway error. Retry with backoff.

Support

Questions? Contact support or email info@thetechcave.net.