AuditLine API reference
The organization's audit log: every write and every refused call, with who asked, from where, and what it committed. On the Scale and Enterprise plans.
Base URL: https://api.withruntime.com. Authentication, idempotency and errors work the same way for every product.
Its states: .
Endpoints#
| Method | Path | Description |
|---|---|---|
GET |
/v2/audit |
List audit |
List audit#
GET /v2/audit
Lists the organization's audit log, newest first, paginated: each write and each refused call, over the days the plan keeps (365 on Scale). On other plans it returns 403.
Its action, for a key's grant: audit:list.
Query:
| Field | Type | Required | Description |
|---|---|---|---|
cursor |
string | No | Where the page starts: a page's next. |
limit |
integer | No | The most items the page holds. At least 1. At most 200. Default: 50. |
action |
string | No | Only lines of this action (keys:create), or of every action of a product (keys:). |
Returns a page: items, each a AuditLine, and next, where the next page starts, none on the last.
In v1: GET /v1/audit.
Examples:
Terminalcurl -X GET "https://api.withruntime.com/v2/audit" \ -H "Authorization: Bearer $RUNTIME_API_KEY"TypeScriptimport { Runtime } from "withruntime";const runtime = new Runtime(); // RUNTIME_API_KEY from the environmentconst result = await runtime.audit.list();Pythonfrom withruntime import Runtimeruntime = Runtime() # RUNTIME_API_KEY from the environmentresult = runtime.audit.list()rubyrequire "withruntime"kit = Withruntime::Generated::Kit.new # RUNTIME_API_KEY from the environmentresult = kit.audit.listTerminalruntime audit listJSON{ "method": "tools/call", "params": { "name": "runtime_audit_list", "arguments": { } } }Objects#
AuditLine
One line of the audit log: a call Access judged, allowed or refused.
| Field | Type | Required | Description |
|---|---|---|---|
id |
string | Yes | Its ID. |
object |
one of audit |
Yes | Always audit. |
action |
string | Yes | What was asked (sandboxes:create, keys:delete). |
allowed |
boolean | Yes | Whether it was allowed. |
reason |
string | No | Why it was refused, or what it changed, when it says. |
target |
string | No | What it acted on: a resource's ID, when it names one. |
accountId |
string | No | The person who asked, or for whom an agent asked. |
agentId |
string | No | The agent or CI pipeline that asked. |
keyId |
string | No | The key it was asked with; none for the console. |
projectId |
string | No | The project it acted in. |
cost |
string (US dollars, decimal) | No | What it committed, in US dollars. |
address |
string | No | The client address it came from. |
requestId |
string | No | The request's ID. |
createdAt |
string (RFC 3339) | Yes | When it was asked. |
Try it on your own work.Signing up takes no card. Your first sandbox asks for one, and the trial starts then. Or paste the setup prompt into your agent.