Relay API Documentation

Connect your existing OpenAI client library by changing just the baseURL and apiKey. Relay transparently proxies the call to the appropriate provider (e.g., Anthropic, DeepSeek, or OpenAI).

curl https://botliy.online/api/v1/chat/completions \
  -H "Authorization: Bearer rly_live_your_relay_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-7-sonnet",
    "messages": [
      {"role": "system", "content": "You are a distributed systems architect."},
      {"role": "user", "content": "Explain raft consensus in two sentences."}
    ]
  }'

Authentication & API Keys

All requests to Relay require authentication using a valid live API key prefixed with rly_live_. Keys are generated in your Dashboard and hashed cryptographically with Argon2id before storing. Relay accepts authentication in two standard header formats:

Connect your frontier terminal coding agents directly to Relay. Seamless native emulation for both Anthropic Messages and OpenAI Chat protocols with zero rate-limit headache and transparent credit billing.

{
  "env": {
    "ANTHROPIC_API_KEY": "YOUR_RELAY_API_KEY",
    "ANTHROPIC_BASE_URL": "https://botliy.online/",
    "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
  },
  "permissions": {
    "allow": [],
    "deny": []
  }
}

claude --version
claude doctor
cd path-to-your-project
claude

Standard chat completion route. Accepts system, user, and assistant messages, alongside temperature, max tokens, and streaming options.

curl https://botliy.online/api/v1/chat/completions \
  -H "Authorization: Bearer rly_live_your_relay_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-7-sonnet",
    "messages": [
      {"role": "system", "content": "You are a distributed systems architect."},
      {"role": "user", "content": "Explain raft consensus in two sentences."}
    ]
  }'

Native Anthropic format supporting anthropic-version: 2023-06-01, structured system messages, multimodal image inputs, and Anthropic SSE streams.

curl https://botliy.online/api/v1/messages \
  -H "x-api-key: rly_live_your_relay_key" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-7-sonnet",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Write an enterprise unit test in Jest"}
    ]
  }'

Direct endpoint for quick single-turn prompt responses, compatible with OpenAI's Responses endpoint interface.

curl https://botliy.online/api/v1/responses \
  -H "Authorization: Bearer rly_live_your_relay_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "input": "Summarize microservice choreography vs orchestration"
  }'

Allows client agents and tools to calculate accurate token weights before sending long prompts or multi-file context windows. This does not consume account credits.

curl https://botliy.online/v1/messages/count_tokens \
  -H "x-api-key: rly_live_your_relay_key" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-7-sonnet",
    "messages": [
      {"role": "user", "content": "Calculate the exact token count before sending"}
    ]
  }'

Returns the list of active public models with context window limits and real-time per-million token pricing in standard OpenAI list format.

{
  "object": "list",
  "data": [
    {
      "id": "claude-3-7-sonnet",
      "object": "model",
      "created": 1715000000,
      "owned_by": "relay",
      "context_window": 200000,
      "pricing": {
        "prompt": 0.000003,
        "completion": 0.000015
      }
    },
    {
      "id": "gpt-4o",
      "object": "model",
      "created": 1715000000,
      "owned_by": "relay",
      "context_window": 128000,
      "pricing": {
        "prompt": 0.0000025,
        "completion": 0.00001
      }
    }
  ]
}

Realtime Server-Sent Events (SSE)

When "stream": true is supplied, Relay establishes an immediate HTTP SSE connection. Upstream tokens are piped directly to your client with negligible latency, while billing meters debit credits asynchronously upon stream completion.

curl https://botliy.online/api/v1/chat/completions \
  -H "Authorization: Bearer rly_live_your_relay_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v3",
    "stream": true,
    "messages": [{"role": "user", "content": "Stream response in realtime token by token"}]
  }'

Smart Routing & Automatic Fallbacks

Relay decouples client-facing model IDs from upstream credentials. When you route through Relay, four intelligent mechanisms ensure maximum uptime:

Rate Limits & Headers

Relay returns explicit rate-limit and correlation headers on every response, ensuring production monitoring systems can track usage:

Error Codes & Taxonomy

Relay uses standard HTTP response codes paired with machine-readable JSON error payloads: