QVeris REST API Documentation

Version: 2026-05-12

The public REST API exposes the core agent path:

Protocol action	Endpoint	Cost behavior
Discover	POST /search	Free; returns ranked capabilities and optional cost signals
Inspect	POST /tools/by-ids	Free; returns full schemas, examples, quality signals, and cost signals
Call	POST /tools/execute	May consume credits according to the selected capability's billing_rule
Usage audit	GET /auth/usage/history/v2	Final request status and charge outcome
Credits ledger	GET /auth/credits/ledger	Final credit balance movements

Replace sample ids such as srch_..., exec_..., and led_... with ids returned by your own API responses.

Base URL

https://qveris.ai/api/v1

Authentication

Send your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Cost and session contract

Discover and Inspect are free. They may return expected_cost, legacy cost, or billing_rule so clients can estimate Call cost before spending credits.

Call can return compact pre-settlement fields such as billing and cost. Final settlement is reported by usage audit and the credits ledger; use those endpoints for support, reconciliation, and user-facing billing history.

session_id is optional. Use one stable value per user task or conversation for tracing, analytics, and pricing context. It is not a cache contract and does not promise cache reuse or session_cache_hit.

Billing transparency contract

QVeris separates pre-call estimate, execution outcome, pre-settlement billing, and final ledger settlement.

Stage	Where to read it	Important fields	How to use it
Pre-call estimate	Discover / Inspect	expected_cost, billing_rule	Show users the pricing rule before executing a capability.
Execution result	Call	success, error_message, execution_outcome	Explain whether the provider/result was usable. Do not use success alone to decide final billing.
Pre-settlement bill	Call and usage audit	billing, pre_settlement_bill, requested_amount_credits	Show the amount requested before final settlement, discounts, or no-charge rules are applied.
Final request status	Usage audit	charge_outcome, reason_code, settlement_result, actual_amount_credits	Answer whether the request was finally charged and why.
Final balance movement	Credits ledger	amount_credits, balance_before, balance_after, execution_id	Reconcile account balance and support tickets.

Recommended reconciliation flow:

1. Use Discover or Inspect to show billing_rule / expected_cost.
2. Call the capability and save execution_id, billing, execution_outcome, and legacy cost.
3. Query /auth/usage/history/v2?execution_id=... to read charge_outcome, reason_code, actual_amount_credits, and credits_ledger_entry_id.
4. Query /auth/credits/ledger or the linked ledger entry to verify the final signed balance movement.

Client guidance:

• REST clients should preserve execution_id, billing, execution_outcome, and cost from Call responses.
• CLI, MCP, and SDK clients should display the same machine-readable fields instead of translating them into unrelated local status names.
• Use stable machine fields such as charge_outcome and reason_code for automation; use billing_summary and error_message for user-facing text.
• cost is kept for backward compatibility. New billing UIs should prefer usage audit and credits ledger for final settlement.

Rate limits

Rate limits are applied per API key when a Bearer token is present, otherwise per client IP.

Action	Default quota
Discover (POST /search)	120 requests/minute
Call (POST /tools/execute)	200 requests/minute

Rate-limited responses include:

Header	Description
X-RateLimit-Limit	Maximum requests allowed in the current window
X-RateLimit-Remaining	Requests remaining in the current window
X-RateLimit-Reset	Unix epoch seconds when the current window resets
Retry-After	Seconds until retry is recommended; always present on 429

1. Discover capabilities

POST /search

Request

{
  "query": "weather forecast API",
  "limit": 10,
  "session_id": "sess_7Q9m"
}

Field	Type	Required	Description
query	string	Yes	Natural-language capability query
limit	integer	No	Maximum result count; default 20, range 1-100
session_id	string	No	Tracking and pricing-context id for this user task

Success response

{
  "query": "weather forecast API",
  "search_id": "srch_01HZX9QK7J3M9T",
  "total": 1,
  "results": [
    {
      "tool_id": "openweathermap.weather.execute.v1",
      "name": "Current Weather",
      "description": "Get current weather data for a city.",
      "provider_name": "OpenWeatherMap",
      "params": [
        {
          "name": "city",
          "type": "string",
          "required": true,
          "description": "City name"
        }
      ],
      "expected_cost": "5 credits per successful request",
      "billing_rule": {
        "unit": "request",
        "amount_credits": 5
      },
      "stats": {
        "avg_execution_time_ms": 210.7,
        "success_rate": 0.982
      }
    }
  ],
  "elapsed_time_ms": 245.6,
  "remaining_credits": 995
}

Response fields

Field	Type	Description
query	string	Original search query when available.
search_id	string	Search id returned by Discover. Use this id in later Inspect or Call requests.
total	integer	Number of capability results returned.
results	array	Ranked capability results.
elapsed_time_ms	number	Search elapsed time in milliseconds.
remaining_credits	number/null	Remaining account credits when available.
error_message	string/null	Error detail for business failures.

Capability result fields

Field	Type	Description
tool_id	string	Unique capability id used by Inspect and Call.
name	string	Human-readable capability name.
description	string	Capability description.
provider_name	string	Capability provider name.
params	array	Parameter definitions. Each item can include name, type, required, description, and enum.
examples	object	Example parameters when available.
expected_cost	string	Human-readable pre-call cost signal when available.
billing_rule	object	Structured cost signal when available.
stats.avg_execution_time_ms	number	Historical average execution time in milliseconds.
stats.success_rate	number	Historical success rate from 0 to 1.

Error responses

Invalid API key:

{
  "query": "weather forecast API",
  "search_id": "srch_failed",
  "total": 0,
  "results": []
}

Insufficient credits:

{
  "query": "weather forecast API",
  "search_id": "srch_failed",
  "total": 0,
  "results": [],
  "error_message": "Insufficient credits",
  "remaining_credits": 0
}

Rate limited:

{
  "status": "failure",
  "status_code": 429,
  "message": "Rate limit exceeded. Please try again later."
}

2. Inspect capabilities by id

POST /tools/by-ids

Inspect returns the same capability result shape as Discover, usually with more complete parameters and examples.

Request

{
  "tool_ids": ["openweathermap.weather.execute.v1"],
  "search_id": "srch_01HZX9QK7J3M9T",
  "session_id": "sess_7Q9m"
}

Field	Type	Required	Description
tool_ids	string[]	Yes	Capability ids returned by Discover
search_id	string	No	Search id that returned the capability
session_id	string	No	Tracking and pricing-context id for this user task

Success response

{
  "search_id": "srch_01HZX9QK7J3M9T",
  "total": 1,
  "results": [
    {
      "tool_id": "openweathermap.weather.execute.v1",
      "name": "Current Weather",
      "description": "Get current weather data for a city.",
      "provider_name": "OpenWeatherMap",
      "params": [
        {
          "name": "city",
          "type": "string",
          "required": true,
          "description": "City name"
        },
        {
          "name": "units",
          "type": "string",
          "required": false,
          "description": "Temperature units",
          "enum": ["metric", "imperial", "standard"]
        }
      ],
      "examples": {
        "sample_parameters": {
          "city": "London",
          "units": "metric"
        }
      },
      "expected_cost": "5 credits per successful request",
      "billing_rule": {
        "unit": "request",
        "amount_credits": 5
      },
      "stats": {
        "avg_execution_time_ms": 210.7,
        "success_rate": 0.982
      }
    }
  ],
  "remaining_credits": 995
}

Response fields

Field	Type	Description
search_id	string	Search id associated with the inspected tools when available.
total	integer	Number of capability results returned.
results	array	Capability results. Each item uses the same capability result fields as Discover.
elapsed_time_ms	number	Inspect elapsed time in milliseconds when available.
remaining_credits	number/null	Remaining account credits when available.
error_message	string/null	Error detail for business failures.

Error responses

Timeout:

{
  "error": "Request timeout",
  "remaining_credits": 995
}

Unexpected proxy failure:

{
  "error": "Tools by-ids failed: upstream service unavailable",
  "remaining_credits": 995
}

3. Call a capability

POST /tools/execute?tool_id={tool_id}

You may pass tool_id as a query parameter or in the JSON body. Use the query parameter form when possible because it is easier to trace in logs.

Request

{
  "search_id": "srch_01HZX9QK7J3M9T",
  "session_id": "sess_7Q9m",
  "parameters": {
    "city": "London",
    "units": "metric"
  },
  "max_response_size": 20480
}

Field	Type	Required	Description
tool_id	string	Required overall	Unique id of the tool to execute. Provide it as the query parameter or in this JSON body.
search_id	string	Recommended	Search id that returned the selected tool
session_id	string	No	Tracking and pricing-context id; if omitted, the service may use the execution id
parameters	object	Yes	Capability-specific parameters from Inspect
max_response_size	integer	No	Truncate long responses; default 20480, -1 disables truncation

Success response

{
  "execution_id": "exec_01HZX9R2R4S2E",
  "result": {
    "data": {
      "temperature": 15.5,
      "description": "partly cloudy"
    }
  },
  "success": true,
  "error_message": null,
  "execution_time": 0.211,
  "elapsed_time_ms": 211,
  "billing": {
    "summary": "5 credits per successful request",
    "list_amount_credits": 5
  },
  "execution_outcome": {
    "outcome": "success",
    "reason_code": "result.valid",
    "provider_success": true,
    "billable_success": true,
    "result_valid": true,
    "user_message": "The provider returned a valid result."
  },
  "cost": 5,
  "remaining_credits": 990
}

Response fields

Field	Type	Description
execution_id	string	Unique id for this execution. Replace sample exec_... values with ids returned by your response.
result	object	Tool result payload. Long responses may use the truncation shape below.
success	boolean	Whether the tool execution succeeded. Do not infer final charge outcome from this field alone.
error_message	string/null	Error detail when success is false.
execution_time	number	Execution elapsed time in seconds. This is the legacy execute response timing field.
elapsed_time_ms	number	Execution elapsed time in milliseconds when available.
billing	object	Compact pre-settlement billing statement when available.
execution_outcome	object	Structured result and billing outcome context when available. Includes stable fields such as outcome, reason_code, provider_success, billable_success, and result_valid.
cost	number	Legacy/pre-settlement cost signal when available.
remaining_credits	number/null	Remaining account credits when available.

Example: empty result, not charged

Some providers return a valid response that contains no usable result. In this case success can be false, the user message should explain the empty result, and the final usage audit should normally classify it as failed_not_charged.

{
  "execution_id": "exec_01HZX9EMPTY",
  "result": {
    "data": {}
  },
  "success": false,
  "error_message": "The provider returned no results for the current parameters. Try different parameters.",
  "execution_time": 0.184,
  "elapsed_time_ms": 184,
  "billing": {
    "summary": "No charge: provider returned no usable result",
    "list_amount_credits": 0
  },
  "execution_outcome": {
    "outcome": "empty_result",
    "reason_code": "result.empty",
    "provider_success": true,
    "billable_success": false,
    "result_valid": false,
    "user_message": "The provider returned no usable result data; this call was not charged."
  },
  "cost": 0,
  "remaining_credits": 990
}

Error responses

Missing tool_id:

{
  "execution_id": "exec_01HZX9R2R4S2E",
  "result": {
    "data": {}
  },
  "success": false,
  "error_message": "Missing required parameter: tool_id. Provide it as query (?tool_id=xxx) or in JSON body.",
  "execution_time": 0.01
}

Insufficient credits:

{
  "execution_id": "exec_01HZX9R2R4S2E",
  "result": {
    "data": {}
  },
  "success": false,
  "error_message": "Insufficient credits",
  "execution_time": 0.01,
  "remaining_credits": 0
}

Upstream tool failure:

{
  "execution_id": "exec_01HZX9R2R4S2E",
  "result": {
    "data": {}
  },
  "success": false,
  "error_message": "Execute API error: HTTP 502",
  "execution_time": 0.211,
  "remaining_credits": 990
}

Long tool responses

If the payload exceeds max_response_size, result may omit data and include truncation fields.

{
  "result": {
    "message": "Result content is too long. Use truncated_content or download full_content_file_url.",
    "full_content_file_url": "https://...",
    "truncated_content": "{\"query\":\"evolution\",\"total_results\":890994",
    "content_schema": {
      "type": "object"
    }
  }
}

Field	Description
truncated_content	Initial bytes of the tool response
full_content_file_url	Temporary URL for the full content
message	LLM-safe explanation of truncation
content_schema	JSON schema for the full content when available

4. Usage audit — final request status

Use usage audit to answer: "Did this request succeed?", "Was a failed request charged?", and "Which execution should support review?" Agent, CLI, and MCP clients should prefer precise filters or summary=true instead of dumping full history into an LLM context.

Endpoint

GET /auth/usage/history/v2

Request headers

Header	Required	Description
Authorization	Yes	Bearer API key

Query parameters

Parameter	Type	Required	Description	Default / range
start_date	string	No	Start of the audit window. Accepts YYYY-MM-DD or ISO-8601 datetime.	-
end_date	string	No	End of the audit window. YYYY-MM-DD expands to the end of that day.	-
event_type	string	No	Exact event type filter: search, search_by_ids, tool_execute, capabilities_query, or model_call.	-
kind	string	No	Higher-level grouping. discover maps to search + search_by_ids; call maps to tool_execute + capabilities_query; model maps to model_call.	-
success	boolean	No	Transport/business success flag recorded for the usage event.	-
billable_success	boolean	No	Billing-specific success flag when available. This can differ from transport success for provider/outcome edge cases.	-
outcome	string	No	Normalized execution outcome filter from execution_outcome.outcome.	-
reason_code	string	No	Normalized execution outcome reason, for example provider or validation reason codes.	-
has_execution_outcome	boolean	No	true returns only events with structured execution outcome; false returns only events without it.	-
charge_outcome	string	No	Final charge classification: charged, included, failed_not_charged, failed_charged_review.	-
anomaly	string	No	Audit anomaly filter: failed_charged_review, missing_ledger_link, missing_billing_snapshot.	-
search_id	string	No	Focus on events linked to a Discover request.	-
execution_id	string	No	Focus on one Call execution. Best filter for "was this call charged?"	-
min_credits	number	No	Minimum effective settled/requested credits. Must be >= 0.	-
max_credits	number	No	Maximum effective settled/requested credits. Must be >= 0.	-
page	integer	No	Page number.	Default 1, minimum 1
page_size	integer	No	Page size when limit is absent.	Default 50, range 1-50000
summary	boolean	No	Include server-side aggregates and high-signal samples. If both dates are omitted, the summary window defaults to the last 24 hours.	Default false
bucket	string	No	Summary time bucket.	hour, day, or week; auto-selects day for windows over 3 days, otherwise hour
limit	integer	No	Overrides returned sample size and clamps it to a context-safe maximum. Use this for Agent/CLI/MCP summaries.	1-50; default summary sample 10

Charge outcome values

Value	Meaning
charged	The effective success flag is true and the settled/effective credit amount is positive.
included	The effective success flag is true and the settled/effective credit amount is zero, for example included credits or a policy exemption.
failed_not_charged	The effective success flag is false and the settled/effective credit amount is zero.
failed_charged_review	The effective success flag is false but the settled/effective amount is positive; treat this as a support/review case.

Common reason codes

reason_code is stable enough for automation, filters, and support workflows. User-facing text can change; machine clients should prefer the code.

Reason code	Typical charge outcome	User-facing meaning
result.valid	charged or included	The provider returned usable data.
result.partial_success	charged, included, or failed_not_charged	The provider returned partial data; inspect the result and billing statement.
result.empty	failed_not_charged	The provider responded but returned no usable result data.
provider.error	failed_not_charged	The provider returned an error.
provider.http_error	failed_not_charged	The provider returned a non-success HTTP response.
provider.rate_limited	failed_not_charged	The upstream provider rate-limited the request.
provider.auth_or_permission	failed_not_charged	The upstream provider rejected auth or permission.
transport.timeout	failed_not_charged	QVeris did not receive a provider response before timeout.
transport.no_response	failed_not_charged	QVeris could not obtain a provider response.
transport.execution_failed	failed_not_charged	The execution path failed before a billable provider result was available.
validation_error	failed_not_charged	Request parameters were invalid or incomplete.
tool_unavailable	failed_not_charged	The selected capability is unavailable.
region_restricted	failed_not_charged	The selected capability is not available in the current region.
oauth_signin_required	failed_not_charged	The capability requires OAuth sign-in before execution.

Example: lookup one execution

curl -sS "$QVERIS_BASE_URL/auth/usage/history/v2?execution_id=exec_01HZX9R2R4S2E" \
  -H "Authorization: Bearer $QVERIS_API_KEY"

{
  "status": "success",
  "message": "Usage events retrieved successfully",
  "status_code": 0,
  "data": {
    "items": [
      {
        "id": "evt_01HZX9R31GH2R",
        "event_type": "tool_execute",
        "source_system": "qveris_website",
        "source_ref_type": "execute_history",
        "source_ref_id": "2b7f7c4a-9f3a-4f61-8b59-3a983a8192a0",
        "session_id": "sess_7Q9m",
        "search_id": "srch_01HZX9QK7J3M9T",
        "execution_id": "exec_01HZX9R2R4S2E",
        "tool_id": "openweathermap.weather.execute.v1",
        "success": true,
        "charge_outcome": "charged",
        "reason_code": "result.valid",
        "duration_ms": 211,
        "billing_snapshot_status": "upstream_provided",
        "billing_rule_snapshot": {
          "unit": "request",
          "amount_credits": 5
        },
        "pre_settlement_bill": {
          "summary": "5 credits per successful request",
          "list_amount_credits": 5
        },
        "settlement_result": {
          "settled_amount_credits": 5
        },
        "pre_settlement_amount_credits": 5,
        "settled_amount_credits": 5,
        "actual_amount_credits": 5,
        "credits_ledger_entry_id": "led_01HZX9R39K6QZ",
        "display_target": "openweathermap.weather.execute.v1",
        "billing_summary": "5 credits per successful request",
        "created_at": "2026-05-16T08:30:12Z"
      }
    ],
    "total": 1,
    "page": 1,
    "page_size": 50,
    "summary": null
  }
}

Example: context-safe summary

curl -sS "$QVERIS_BASE_URL/auth/usage/history/v2?summary=true&bucket=day&kind=call&limit=5&start_date=2026-05-01&end_date=2026-05-16" \
  -H "Authorization: Bearer $QVERIS_API_KEY"

{
  "status": "success",
  "message": "Usage events retrieved successfully",
  "status_code": 0,
  "data": {
    "items": [
      {
        "id": "evt_01HZX9R31GH2R",
        "event_type": "tool_execute",
        "execution_id": "exec_01HZX9R2R4S2E",
        "tool_id": "openweathermap.weather.execute.v1",
        "success": true,
        "charge_outcome": "charged",
        "settled_amount_credits": 5,
        "created_at": "2026-05-16T08:30:12Z"
      }
    ],
    "total": 42,
    "page": 1,
    "page_size": 5,
    "summary": {
      "start_date": "2026-05-01T00:00:00Z",
      "end_date": "2026-05-16T23:59:59.999999Z",
      "bucket": "day",
      "total_count": 42,
      "success_count": 40,
      "failure_count": 2,
      "charge_outcome_counts": {
        "charged": 35,
        "included": 5,
        "failed_not_charged": 2,
        "failed_charged_review": 0
      },
      "pre_settlement_credits": 210,
      "settled_credits": 175,
      "max_charge_items": [],
      "buckets": [
        {
          "bucket_start": "2026-05-16T00:00:00Z",
          "total_count": 8,
          "success_count": 8,
          "failure_count": 0,
          "charged_count": 7,
          "included_count": 1,
          "failed_not_charged_count": 0,
          "failed_charged_review_count": 0,
          "pre_settlement_credits": 40,
          "settled_credits": 35
        }
      ]
    }
  }
}

Response fields

Top-level response uses the standard APIResponse envelope.

Field	Type	Description
status	string	success or failure.
message	string	Human-readable server message.
status_code	integer	Application status code. Success is 0; validation failures use negative codes.
data.items	array	Usage events in reverse chronological order.
data.total	integer	Total rows matching filters.
data.page	integer	Current page.
data.page_size	integer	Effective returned item/sample size. If limit is set, it overrides page_size and is capped at 50.
data.summary	object/null	Aggregate summary when summary=true; otherwise null.

Important data.items[] fields:

Field	Description
event_type	Canonical event type. search = Discover, search_by_ids = Inspect, tool_execute / capabilities_query = Call, model_call = model usage.
search_id / execution_id	Correlation ids for the Discover or Call flow.
success	Recorded success flag for the usage event.
charge_outcome	User-facing final charge classification. Use this instead of guessing from success alone.
error_message	Error details when available.
duration_ms	Request duration in milliseconds.
request_payload / response_payload_summary	Stored request/response summaries for audit. Agent clients should avoid dumping these by default.
execution_outcome and outcome fields	Structured provider/result outcome details when available.
billing_rule_snapshot	Billing rule captured at request time.
pre_settlement_bill	Pre-settlement billing statement captured before final ledger settlement.
settlement_result	Final settlement details when available.
requested_amount_credits / actual_amount_credits	Requested versus settled/effective credits.
credits_ledger_entry_id	Ledger row id when this usage event produced a final balance movement.
display_target / billing_summary	UI-safe target and billing summary.

Error responses

Invalid date or bucket:

{
  "status": "failure",
  "message": "Invalid start_date format. Use YYYY-MM-DD or ISO-8601 datetime",
  "status_code": -7,
  "data": null
}

Invalid credit range:

{
  "status": "failure",
  "message": "min_credits cannot be greater than max_credits",
  "status_code": -7,
  "data": null
}

5. Credits ledger — final balance movements

Use the credits ledger to explain the final account balance. Usage audit describes requests; the ledger describes immutable credit movements. A charged Call should normally have a usage event with charge_outcome=charged and a linked ledger item.

Endpoint

GET /auth/credits/ledger

Request headers

Header	Required	Description
Authorization	Yes	Bearer API key

Query parameters

Parameter	Type	Required	Description	Default / range
start_date	string	No	Start of the ledger window. Accepts YYYY-MM-DD or ISO-8601 datetime.	-
end_date	string	No	End of the ledger window. YYYY-MM-DD expands to the end of that day.	-
entry_type	string	No	Exact ledger event type, for example consume_tool_execute.	-
scope	string	No	Preset entry-type group. account_history includes grant_payment_recharge, consume_tool_search, consume_tool_execute, and consume_model_call.	-
direction	string	No	Balance direction. consume returns negative credit movements; grant returns positive movements; any returns both.	Default any; allowed consume, grant, any
min_credits	number	No	Minimum absolute credit amount. For example min_credits=5 matches both -5 and +5. Must be >= 0.	-
max_credits	number	No	Maximum absolute credit amount. Must be >= 0.	-
page	integer	No	Page number.	Default 1, minimum 1
page_size	integer	No	Page size when limit is absent.	Default 50, range 1-500
summary	boolean	No	Include aggregate balance movement summary. If both dates are omitted, the summary window defaults to the last 24 hours.	Default false
bucket	string	No	Summary time bucket.	hour, day, or week; auto-selects day for windows over 3 days, otherwise hour
limit	integer	No	Overrides returned sample size and summary max-amount samples, capped for Agent/CLI/MCP use.	1-50; default summary sample 10

Common entry_type values

Value	Meaning
grant_payment_recharge	Credits granted by a recharge/payment.
grant_welcome_bonus	Welcome or promotional credit grant.
grant_invitation_reward	Invitation/referral credit grant.
consume_tool_search	Credits consumed for Discover when a deployment charges search.
consume_tool_execute	Credits consumed for a capability Call.
consume_model_call	Credits consumed for model calls.
consume_payment_refund	Credit movement related to a payment refund.

Example: recent Call charges

curl -sS "$QVERIS_BASE_URL/auth/credits/ledger?entry_type=consume_tool_execute&page=1&page_size=10" \
  -H "Authorization: Bearer $QVERIS_API_KEY"

{
  "status": "success",
  "message": "Credits ledger retrieved successfully",
  "status_code": 0,
  "data": {
    "items": [
      {
        "id": "led_01HZX9R39K6QZ",
        "entry_type": "consume_tool_execute",
        "amount_credits": -5,
        "source_system": "qveris_website",
        "source_ref_type": "execute_history",
        "source_ref_id": "2b7f7c4a-9f3a-4f61-8b59-3a983a8192a0",
        "execution_id": "exec_01HZX9R2R4S2E",
        "pre_settlement_bill": {
          "execution_id": "exec_01HZX9R2R4S2E",
          "summary": "5 credits per successful request",
          "list_amount_credits": 5
        },
        "settlement_result": {
          "settled_amount_credits": 5
        },
        "balance_before": {
          "total_available_credits": 995
        },
        "balance_after": {
          "total_available_credits": 990
        },
        "description": "Tool execution charge",
        "created_at": "2026-05-16T08:30:13Z"
      }
    ],
    "total": 1,
    "page": 1,
    "page_size": 10,
    "summary": null
  }
}

Example: aggregate balance movements

curl -sS "$QVERIS_BASE_URL/auth/credits/ledger?summary=true&scope=account_history&direction=any&bucket=day&limit=5&start_date=2026-05-01&end_date=2026-05-16" \
  -H "Authorization: Bearer $QVERIS_API_KEY"

{
  "status": "success",
  "message": "Credits ledger retrieved successfully",
  "status_code": 0,
  "data": {
    "items": [
      {
        "id": "led_01HZX9R39K6QZ",
        "entry_type": "consume_tool_execute",
        "amount_credits": -5,
        "source_ref_type": "execute_history",
        "source_ref_id": "2b7f7c4a-9f3a-4f61-8b59-3a983a8192a0",
        "execution_id": "exec_01HZX9R2R4S2E",
        "created_at": "2026-05-16T08:30:13Z"
      }
    ],
    "total": 18,
    "page": 1,
    "page_size": 5,
    "summary": {
      "start_date": "2026-05-01T00:00:00",
      "end_date": "2026-05-16T23:59:59.999999",
      "bucket": "day",
      "total_entries": 18,
      "consume_count": 14,
      "grant_count": 4,
      "consumed_credits": 175,
      "granted_credits": 1000,
      "net_amount_credits": 825,
      "max_amount_items": [],
      "buckets": [
        {
          "bucket_start": "2026-05-16T00:00:00",
          "entry_count": 3,
          "consume_count": 3,
          "grant_count": 0,
          "consumed_credits": 15,
          "granted_credits": 0,
          "net_amount_credits": -15
        }
      ]
    }
  }
}

Response fields

Field	Type	Description
data.items	array	Ledger rows in reverse chronological order.
data.total	integer	Total rows matching filters.
data.page / data.page_size	integer	Current page and effective returned item/sample size.
data.summary	object/null	Aggregate balance summary when summary=true; otherwise null.

Important data.items[] fields:

Field	Description
entry_type	Immutable ledger event type.
amount_credits	Signed balance movement. Negative values consume credits; positive values grant credits.
source_system	System that created the ledger row.
source_ref_type / source_ref_id	Source row reference for backend audit.
execution_id	Call execution id returned by /tools/execute; use this for user reconciliation. Present for Call ledger rows when available.
pre_settlement_bill	Billing snapshot before final settlement.
settlement_result	Final settlement result.
balance_before / balance_after	Balance snapshots around this movement when available.
ledger_metadata	Additional internal-safe metadata for audit/debugging.
description	Human-readable ledger description.
created_at	Creation timestamp.

Summary fields:

Field	Description
total_entries	Count of matching ledger rows.
consume_count / grant_count	Number of negative and positive movements.
consumed_credits / granted_credits	Absolute consumed and granted totals.
net_amount_credits	Signed net sum; grants positive, consumption negative.
max_amount_items	High-signal largest absolute movements, capped by limit.
buckets	Per-bucket time series for charts or compact Agent summaries.

Error responses

Invalid direction:

{
  "status": "failure",
  "message": "Invalid direction. Use consume, grant, or any",
  "status_code": -7,
  "data": null
}

Invalid credit range:

{
  "status": "failure",
  "message": "min_credits must be greater than or equal to 0",
  "status_code": -7,
  "data": null
}

End-to-end smoke checklist

1. Create a fresh session_id.
2. Run Discover and save search_id.
3. Inspect the selected tool_id; confirm required params and pre-call cost fields.
4. Call with valid parameters; save execution_id.
5. Query usage audit by execution_id.
6. Query the credits ledger and confirm the final balance movement matches the audit outcome.

OpenAPI

The public OpenAPI document is available in this repository at docs/openapi/qveris-public-api.openapi.json. It includes request bodies, response schemas, and examples for the Discover, Inspect, and Call path.