moonming · May 18, 2026 · May 15, 2026 · May 16, 2026 · May 16, 2026 · May 16, 2026
diff --git a/docs/configuration/api-keys.md b/docs/configuration/api-keys.md
@@ -108,9 +108,6 @@ Managed budget enforcement exists on the managed `/dp/budget_check` path.
 Current standalone boundary:
 
 - standalone self-hosted deployments default to a disabled budget client, which is allow-all
-- the standalone admin write validator currently rejects `max_budget_usd`, even though broader typed and OpenAPI surfaces reference it
-
-Do not treat `max_budget_usd` as part of the current verified standalone admin write contract.
 
 ## Troubleshooting
 

diff --git a/docs/configuration/budgets.md b/docs/configuration/budgets.md
@@ -6,8 +6,6 @@ sidebar_position: 37
 
 Budget enforcement in the current gateway runtime is driven by the managed budget-check path, not by a standalone in-process budget engine.
 
-Use this page to avoid over-assuming what `max_budget_usd` means in standalone deployments.
-
 ## Current Runtime Model
 
 Before dispatch, the proxy can call:
@@ -27,13 +25,6 @@ Current boundary:
 - managed deployments can attach a live budget client through the managed data-plane path
 - standalone self-hosted deployments default to `BudgetClient::disabled()`, which allows requests through
 
-Because of that boundary, `max_budget_usd` should not be treated as part of the current verified standalone admin write contract.
-
-Current standalone caveat:
-
-- the typed `ApiKey` model and admin OpenAPI mention `max_budget_usd`
-- the active standalone admin JSON Schema validator rejects `max_budget_usd` on write today
-
 ## Operator Guidance
 
 - treat managed mode as the real budget-enforcement path today
@@ -61,11 +52,9 @@ This is a caller-visible denial, not just an internal accounting event.
 
 Inspect budget-check freshness and the cached-decision behavior first.
 
-### A standalone deployment ignores `max_budget_usd`
-
-That is expected with the current standalone runtime boundary.
+### A standalone deployment does not enforce per-key budgets
 
-If you are using the standalone admin API directly, do not send `max_budget_usd` until the write validator and runtime contract are aligned.
+That is expected with the current standalone runtime boundary. Standalone deployments default to allow-all; per-key budget enforcement lives in the managed control plane via the budget-check path.
 
 ## Related Pages
 

diff --git a/docs/integration/errors-and-retries.md b/docs/integration/errors-and-retries.md
@@ -40,7 +40,7 @@ The proxy emits the following gateway-generated failures. The `error.type` strin
 | `413`  | `invalid_request_error`  | Request body exceeded `proxy.request_body_limit_bytes` |
 | `422`  | `content_filter`         | Guardrail blocked the request or response content |
 | `429`  | `rate_limit_exceeded`    | Rate-limit rejection (per-key or per-model) |
-| `429`  | `billing_error` (with `code: "budget_exceeded"`) | Budget rejection from the ApiKey's `max_budget_usd` |
+| `429`  | `billing_error` (with `code: "budget_exceeded"`) | Budget rejection from the managed budget controller |
 | `503`  | `provider_unavailable`   | No provider bridge is registered for the resolved provider |
 | `503`  | `all_candidates_unavailable` | Every routing candidate was excluded by runtime status (cooldown or background-unhealthy) and the routing model is configured with `on_all_filtered: fail`. The response carries `Retry-After: 30`. See [Routing And Failover § All-Targets-Filtered Policy](../configuration/routing-and-failover.md#all-targets-filtered-policy) |
 

diff --git a/docs/overview/core-concepts.md b/docs/overview/core-concepts.md
@@ -53,8 +53,6 @@ An API key also carries:
 
 `team_id` and `owner_id` are bucket identifiers consumed by `team`-scoped and `member`-scoped [`RateLimitPolicy`](#rate-limit-policy) rows. They are not access controls on their own.
 
-Managed budget paths may also reference per-key budget state, but `max_budget_usd` is not part of the current verified standalone admin write contract.
-
 An empty `allowed_models` list denies access to every model. A wildcard entry `"*"` allows access to every model in scope.
 
 ## Rate Limit Policy

diff --git a/docs/overview/feature-matrix.md b/docs/overview/feature-matrix.md
@@ -25,7 +25,7 @@ Use it as a navigation aid, not as a replacement for detailed feature pages.
 | Provider-specific passthrough | Available | Use `/passthrough/:provider/*rest` for unsupported or provider-native routes. |
 | Standalone admin API | Available | Current admin surface includes models, API keys, provider keys, guardrails, cache policies, observability exporters, health, metrics, OpenAPI, and playground. |
 | API key allowlist authz | Available | Uses hashed caller keys and model allowlists. |
-| Per-key budgets | Limited | Live enforcement is currently centered on managed-mode `/dp/budget_check`. Standalone self-hosted mode defaults to allow-all, and the standalone admin write validator does not currently accept `max_budget_usd`. |
+| Per-key budgets | Limited | standalone deployments do not enforce per-key budgets; managed deployments enforce via the budget controller. |
 | Rate limits and concurrency limits | Available | Three layers are AND-combined per request: `ApiKey.rate_limit`, `Model.rate_limit`, and scope-matched `RateLimitPolicy` rows (`api_key` / `model` / `team` / `member`). |
 | Routing models and failover | Available | Current model schema supports routing strategies and retry budget behavior. |
 | Keyword guardrails | Available | Current runtime enforcement is on `POST /v1/chat/completions`; non-chat endpoints do not run the guardrail chain today. |

diff --git a/docs/reference/resource-schemas.md b/docs/reference/resource-schemas.md
@@ -42,7 +42,6 @@ Not every field or shape present in the schema should be interpreted as equally
 
 Examples:
 
-- the typed `ApiKey` model and admin OpenAPI mention `max_budget_usd`, but the current standalone admin write validator rejects it and standalone hard-stop budget behavior is not the current documented default
 - `Model.rate_limit` and `ApiKey.rate_limit` are both enforced today, alongside scope-matched `RateLimitPolicy` rows. See [Rate Limits](../configuration/rate-limits.md) for the layer order and the AND-combination semantics.
 - `Model.background_model_check` exists in schema, but it only applies to direct models and its runtime effect is surfaced through `/admin/v1/models/status`
 - `Guardrail.kind = bedrock` exists in schema, but current generally reliable runtime behavior is strongest on `keyword`