Costs API¶
Soulacy records LLM token usage for every agent invocation. Estimated cost_usd is populated when costs.pricing is configured for the provider/model; unknown prices remain 0 rather than guessed. Use this API to monitor spend by agent, user, or time period.
Cost summary¶
Query parameters¶
| Parameter | Default | Description |
|---|---|---|
period |
30d |
Time period: 1d, 7d, 30d, 90d, or all |
agent_id |
— | Filter by agent ID |
user_id |
— | Filter by user ID |
Response¶
{
"period": "30d",
"total_cost_usd": 4.27,
"total_tokens": {
"input": 1240000,
"output": 320000
},
"by_provider": {
"openai": {
"cost_usd": 3.92,
"input_tokens": 1100000,
"output_tokens": 300000
},
"anthropic": {
"cost_usd": 0.35,
"input_tokens": 140000,
"output_tokens": 20000
}
}
}
Per-agent breakdown¶
Query parameters¶
Same as summary (period, agent_id, user_id).
Response¶
{
"period": "30d",
"agents": [
{
"agent_id": "researcher",
"model": "gpt-4o",
"invocations": 142,
"cost_usd": 3.15,
"input_tokens": 890000,
"output_tokens": 210000
},
{
"agent_id": "assistant",
"model": "gpt-4o-mini",
"invocations": 1203,
"cost_usd": 1.12,
"input_tokens": 350000,
"output_tokens": 110000
}
]
}
Per-invocation log¶
Query parameters¶
| Parameter | Default | Description |
|---|---|---|
limit |
50 |
Records per page |
offset |
0 |
Pagination offset |
agent_id |
— | Filter by agent |
since |
— | ISO 8601 timestamp |