Docs
Docs

Syllogy API v1

New users receive promo credits while the promotion is active. Bitcoin Lightning top‑ups are available now via admin endpoints. Track your activity on the Usage page.
Request content—including prompts, system instructions, message history, tool payloads, and model outputs—is processed for the request and not retained by Syllogy. Aggregate token, credit, latency, and status metrics are retained.

All endpoints use JSON over HTTPS.

Management endpoints require an admin key (your session key or a special admin API key). Keys you create via the API are non-admin and are valid for /api/v1/chat/completions.

Models

GET/api/v1/modelsList available model IDs (OpenAI-compatible format)
Response example 
{
  "object": "list",
  "data": [ { "id": "mistral-7b", "object": "model" } ]
}

Auth

POST/api/v1/auth/otpSend one-time code to email
Body {"email":"user@example.com"}
POST/api/v1/auth/loginExchange OTP ⇢ session / API key
Body {"email":"…","otp":"123456","type":"web|api"}

Chat Completions

POST/api/v1/chat/completionsNon-streaming or streaming (SSE) responses via request body
Request schema 
{
  "model": "model-id",         // required; see GET /api/v1/models
  "messages": [                 // ≤ 500 items
    { "role": "system", "content": "…" },
    { "role": "user",   "content": "…" }
  ],
  "stream": true,               // optional; set true for SSE stream
  "max_tokens": 256,            // optional; if omitted, server allows the
                                 // maximum generation tokens the model can
                                 // fit given the current prompt/context
  "tools": […],                 // optional; OpenAI tool schema
  "tool_choice": …              // optional; tool selection
}
curl examples 
# Non-streaming
curl -s \
  -H "Authorization: Bearer $SYLLOGY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "model":"mistral-7b",
        "messages":[{"role":"user","content":"Hello"}]
      }' \
  https://syllogy.ai/api/v1/chat/completions | jq -r '.choices[0].message.content'

# Streaming (SSE)
curl -sN \
  -H "Authorization: Bearer $SYLLOGY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "model":"mistral-7b",
        "stream":true,
        "messages":[{"role":"user","content":"Write a haiku"}]
      }' \
  https://syllogy.ai/api/v1/chat/completions | sed -n 's/^data: //p'
Streaming uses text/event-stream with data: {json} lines ending in [DONE], matching OpenAI SSE format.

Account

GET/api/v1/account/infoEmail, credits, timestamps (admin only)
POST/api/v1/account/logoutInvalidate current session / delete current key (admin only)

API Key Management

GET/api/v1/keys/listList your API keys (admin only)
POST/api/v1/keys/newCreate new non-admin API key (admin only)
POST/api/v1/keys/deleteDelete key  Body {"key_hash":"…"} (admin only)

Lightning Invoices

POST/api/v1/invoice/getGet or create a 0‑amount BOLT11 invoice for the admin key’s user (admin only)
No body → Response {"r_hash":"…","bolt11":"lnbc…","expires_at":1721000000}
GET/api/v1/invoice/list?offset=0List invoices (paginated, max 200 per page) (admin only)
Response [{"r_hash":"…","sats":123,"settled_at":1721000000}, …]
Invoice semantics
  • Only one pending invoice (msats=0) exists per user email at a time. Repeated /invoices/get returns the same BOLT11 until it’s paid or expires.
  • Background polling settles paid invoices (sets msats, settled_at) and credits the user with msats / 1000 (floored). Expired, unpaid invoices are deleted.
  • /invoices/list returns sats as floor(msats/1000) and settled_at as a Unix timestamp. Use offset for pagination.
curl examples (admin key) 
# Get or create a 0-amount invoice (derived from the admin key's user)
curl -s \
  -H "Authorization: Bearer $ADMIN_API_KEY" \
  -X POST \
  https://syllogy.ai/api/v1/invoice/get | jq

# List invoices (200 per page), offset 0
curl -s \
  -H "Authorization: Bearer $ADMIN_API_KEY" \
  'https://syllogy.ai/api/v1/invoice/list?offset=0' | jq

Rate limit: 1 req/sec per IP plus a global cap. Exceeding returns HTTP 429.