> ## Documentation Index > Fetch the complete documentation index at: https://docs.vumasign.com/llms.txt > Use this file to discover all available pages before exploring further. # One endpoint, with its health. > ⚠️ THREE SITUATIONS ANSWER WITH THE SAME 404: no such endpoint; one belonging to another organisation; an id that is not a uuid. "This exists but is not yours" is itself the secret, and the handler could not distinguish them if it wanted to — the read runs inside a transaction already scoped to the key’s tenant. ## OpenAPI ````yaml /openapi.json get /api/v1/webhooks/{webhookId} openapi: 3.1.0 info: title: Vumasign API version: 0.1.0 description: >- The Vumasign API. South African e-signature: templates, the addresses an integration writes values to, an idempotent send, and webhooks. **This document describes what exists today and nothing else.** It is not a roadmap, and there is deliberately no entry here for anything that has not shipped — a spec you can generate a client from and then discover was aspirational is worse than no spec at all. ## Authentication Authorization: Bearer vsk_live_… One header, one step. Keys are minted in Settings. `vsk_live_` and `vsk_test_` are different keys; the prefix is stored in clear so a leaked key is greppable and traceable to an organisation, while the secret is stored only as a hash. ## Errors Every refusal the API returns deliberately, at every status, is `{ "error": { "code", "message" } }`, and such a response is a failure **if and only if** it carries an `error` key — that test is correct even for a client that ignores the status line. An unexpected server failure may not have this shape: treat a 5xx without an `error` key as an `internal_error`, which is retryable. `code` is the contract and is stable forever; `message` is prose for a developer reading a log and may be reworded in any release. The root of this document carries **`x-error-codes`**: the closed enum, with `status`, `retryable`, `backoff`, `emitted` and `guidance` for each code. Branch on that rather than on our prose. ## Rate limits **600 requests a minute on any paid plan, 60 on the free plan — and a test key gets the same number as a live one.** Per key, not per organisation, so a batch job and an interactive integration can be given a key each and neither can starve the other. The ceiling follows the PLAN rather than the kind of key, because the plan is what says whether this is production traffic or somebody evaluating us. On the API plan it is sized so that no honest workload meets it: a thousand envelopes with a status read each is two thousand requests, which is three and a half minutes rather than seventeen. The sandbox deliberately gets the *same* number rather than a larger one — an integration that passes in rehearsal is then an integration that passes live, and your `429` handling is exercised where meeting a `429` is free. Every successful authenticated response carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset` (seconds until the window closes, at most 60). Read them on your successes and you will never need to be refused. A 429 carries them too, and additionally carries `Retry-After`, which you should honour. A 401 carries none of them; a 403 may not, depending on whether it was decided before the request was counted. Each operation's responses say which others do. ⚠️ **A refused request still counts.** Retrying inside the window pushes the counter higher rather than holding it at the ceiling — which is why the backoff in `x-error-codes` is exponential and why `Retry-After` is worth reading rather than guessing at. ⚠️ The limits are **per minute**, and the window is fixed rather than sliding. Spending a whole window at the end of one minute and a whole window at the start of the next is possible; it is bounded at two windows' worth, which is a burst rather than an incident. ## Webhooks Register an endpoint with `POST /api/v1/webhooks`; the response carries the signing secret, **once**. The root of this document carries **`x-webhooks`**: the body we POST, the signature recipe (including the two mistakes that make a verifier look correct and be worthless), the published retry ladder, and the auto-disable rule. Read that before writing a receiver. ## What is not here yet No organisation-level email exists, so an auto-disabled webhook endpoint is discovered by polling rather than by being told. There is no way to change an envelope after it is created and before it is sent — `PATCH /v1/envelopes/{id}` is designed and not built — so a draft is created complete or not at all. Both are stated because a gap named is a gap somebody can plan around. ⚠️ **The rate limit above counts authenticated requests only.** A request carrying no key, or a key we cannot recognise, is refused before any counter is touched — a per-key ceiling cannot defend against a caller who has not presented a key, and pretending otherwise would be a promise we could not keep. That half belongs to the edge rather than to this API. (This section has twice described as missing something that now exists: the `uri` on every webhook body, which was `null` until `GET /api/v1/envelopes/{envelopeId}` gave it somewhere to point, and the rate limiter, which nothing enforced until the per-key ceiling above. Both were named here as gaps first, which is the point of naming them.) servers: - url: https://app.vumasign.com security: - ApiKey: [] tags: - name: Templates description: >- A template is a document with its signing roles and field placements authored once. Every envelope is made from one, or from a one-off pack. - name: Documents description: >- The pages of a template’s documents, served as PDF so that a client which has to show the paper can do so without rendering it itself. - name: Envelopes description: >- An envelope is one sending of a document to named people. Creating one is not sending it: a draft emails nobody until it is sent. - name: Webhooks description: >- Endpoints we POST signed events to when an envelope changes, with the health of each endpoint and a secret you can rotate without a gap. - name: Specification description: >- This document, served without a credential so that anybody can read what the API does before they hold a key. paths: /api/v1/webhooks/{webhookId}: get: tags: - Webhooks summary: One endpoint, with its health. description: >- ⚠️ THREE SITUATIONS ANSWER WITH THE SAME 404: no such endpoint; one belonging to another organisation; an id that is not a uuid. "This exists but is not yours" is itself the secret, and the handler could not distinguish them if it wanted to — the read runs inside a transaction already scoped to the key’s tenant. operationId: getWebhook parameters: - name: webhookId in: path required: true description: The endpoint’s `id`, as returned by the list. schema: type: string format: uuid examples: - 01960000-0000-4000-8000-000000000e5d description: An endpoint id. responses: '200': description: The endpoint and its delivery health, without its secret. content: application/json: schema: $ref: '#/components/schemas/WebhookEndpoint' description: The endpoint. examples: example: summary: One healthy endpoint value: id: 01960000-0000-4000-8000-000000000e5d url: https://api.example.test/hooks/vumasign subscribed_events: - envelope.completed - envelope.declined brand_id: null active: true consecutive_failures: 0 last_success_at: '2026-09-02T09:06:00.000Z' last_failure_at: null disabled_at: null disabled_reason: null created_at: '2026-09-01T08:00:00.000Z' headers: RateLimit-Limit: description: >- How many requests this key may make per minute: 600 on any paid plan and 60 on the free plan, the same in both environments. Per key, not per organisation — a second key has its own allowance. ⚠️ READ IT RATHER THAN ASSUMING IT: it is the number the database applied to this request, so it is right even when this description is out of date. schema: type: integer description: Requests permitted in the current minute. RateLimit-Remaining: description: >- How many of them are left, this one already counted. Zero on the request that is refused, and zero on every further request of that minute — ⚠️ a refused request still counts. schema: type: integer description: Requests remaining in the current minute. RateLimit-Reset: description: >- Seconds until the current minute closes and the allowance returns. At most 60, which is what makes the published exponential backoff converge in about six doublings rather than eleven. schema: type: integer description: Seconds until the window resets. '401': description: >- `unauthenticated` — The credential was absent, unreadable, unknown, wrong or revoked — this answer is deliberately identical for all of them, so it cannot be used to probe which keys exist. Retrying the same request changes nothing. Check the key in Settings or issue a new one. content: application/json: schema: $ref: '#/components/schemas/Error' description: The refusal. '403': description: >- `forbidden` — The key is valid and the organisation’s plan does not grant what this operation does. Nothing about the credential needs to change, and it is deliberately MORE specific than 401 because the caller has already proved they hold the key. TWO SITUATIONS PRODUCE IT. (1) THE PLAN EXCLUDES THE OPERATION: `POST /api/v1/templates` on the Free plan, which does not include authoring templates. A template that already exists stays readable at `GET /api/v1/templates`; authoring one needs a plan that includes templates. (2) WE CANNOT ESTABLISH WHAT THE PLAN INCLUDES, on any operation, so nothing is granted. That one is ours rather than yours, an ordinary account does not meet it, and it means ask us. `insufficient_scope` — This key is scoped and does not hold the scope this operation requires — `x-required-scope` on the operation names it, and the `WWW-Authenticate` header on the refusal repeats it (RFC 6750 §3.1). ⚠️ Retrying cannot help and neither can editing the key: scopes are fixed when a key is minted and no endpoint or screen can widen them. Issue a new key with the access it needs. A key created before scopes existed carries an empty scope list, which means EVERY capability, so this refusal cannot reach an integration that was already working. content: application/json: schema: $ref: '#/components/schemas/Error' description: The refusal. '404': description: >- `not_found` — No such resource — or none this key’s organisation can see, or one that has been archived, or an id that is not a uuid. Those four are one answer on purpose: "it exists but is not yours" is itself a disclosure. Do not retry; check the id against the list endpoint. content: application/json: schema: $ref: '#/components/schemas/Error' description: The refusal. '429': description: >- `rate_limited` — This key has spent its minute. The limit is per KEY and per minute, and it follows your PLAN rather than the kind of key — 600 requests a minute on every paid plan, 60 on the free plan, and the sandbox gets the same number as production so that an integration which passes in rehearsal passes live. Every successful authenticated response, and this one, carries `RateLimit-Limit`, `RateLimit-Remaining` and `RateLimit-Reset` (seconds), so a client can slow down before it is made to. TWO OPERATIONS ALSO ANSWER IT FOR A REASON OF THEIR OWN. `POST /api/v1/envelopes/batches` costs one request per row and is refused WHOLE when its rows do not fit what is left of the key’s minute — `RateLimit-Remaining` says how many would. `GET /api/v1/envelopes/{envelopeId}/documents/{position}/draft` has its own per-envelope and per-organisation ceilings on renders each minute, and the `message` says which one refused; there `Retry-After` is that ceiling’s, while the `RateLimit-*` headers still describe the key. `Retry-After` on this refusal is at most 60 and is read from the stored window rather than computed, so two of our instances refusing the same key quote the same instant. ⚠️ A REFUSED REQUEST STILL COUNTS: hammering a limit you have already exceeded pushes the counter higher rather than holding it. Honour `Retry-After`. If one key genuinely needs more throughput, issue a second one in Settings — the ceiling is a fairness and blast-radius control, not a commercial meter. content: application/json: schema: $ref: '#/components/schemas/Error' description: The refusal. headers: RateLimit-Limit: description: >- How many requests this key may make per minute: 600 on any paid plan and 60 on the free plan, the same in both environments. Per key, not per organisation — a second key has its own allowance. ⚠️ READ IT RATHER THAN ASSUMING IT: it is the number the database applied to this request, so it is right even when this description is out of date. schema: type: integer description: Requests permitted in the current minute. RateLimit-Remaining: description: >- How many of them are left, this one already counted. Zero on the request that is refused, and zero on every further request of that minute — ⚠️ a refused request still counts. schema: type: integer description: Requests remaining in the current minute. RateLimit-Reset: description: >- Seconds until the current minute closes and the allowance returns. At most 60, which is what makes the published exponential backoff converge in about six doublings rather than eleven. schema: type: integer description: Seconds until the window resets. Retry-After: description: >- Seconds to wait before retrying, from the stored end of the window. Always present on a 429, and never more than 60. On the per-key limit it equals `RateLimit-Reset`; a narrower limit, such as the draft render allowance, has a window of its own and may ask for a different wait. `x-error-codes` tells you to honour it; this is it. schema: type: integer description: Seconds to wait before retrying. '500': description: >- `internal_error` — Ours, not yours. ⚠️ RETRY WITH THE SAME `Idempotency-Key` YOU SENT THE FIRST TIME, on any operation that takes one — that is what makes a retry safe here, and sending a NEW key would create a second envelope or send the same one twice. The safe GETs in this document can be retried freely. (This entry used to say every operation here was a GET; the write endpoints have since shipped, and the reasoning moved with them.) content: application/json: schema: $ref: '#/components/schemas/Error' description: The refusal. security: - ApiKey: [] components: schemas: WebhookEndpoint: type: object description: >- A registered endpoint: where to deliver, what to deliver, and how healthy it is. ⚠️ CARRIES NO `secret`. The secret is returned exactly twice in this API — by the registration and by a rotation — and is never readable afterwards. properties: id: type: string format: uuid description: The endpoint’s id. A bare uuid. url: type: string description: >- Where deliveries are POSTed. `https://` only, no credentials, and not a private or loopback address — deliveries are made from inside our network. subscribed_events: type: array description: >- What this endpoint hears about. Never empty. ⚠️ An unknown name is refused at registration rather than accepted and never delivered. items: type: string description: One event type. enum: - envelope.sent - envelope.completed - envelope.declined - envelope.voided - envelope.expired - envelope.sealed - envelope.sealing_failed - recipient.completed brand_id: type: - string - 'null' format: uuid description: >- Deliver only envelopes carrying this brand; null for every envelope in the organisation. ⚠️ **THIS IS ROUTING, NOT ISOLATION.** Any key on this organisation can read every envelope in it whatever brand it carries — there is no brand predicate in any access rule anywhere. Filtering decides which events are POSTed to which URL; it does not, and cannot, stop a consumer learning about another brand by asking. Do not build a permission boundary out of it. active: type: boolean description: >- Whether we are delivering. False either because you deactivated it or because it was auto-disabled after failing for the whole health window — `disabled_reason` says which, in prose. consecutive_failures: type: integer description: >- ⚠️ DELIVERIES THAT EXHAUSTED THE WHOLE LADDER SINCE THE LAST SUCCESS, not individual attempts. One unreachable host over one event counts once here, not seven times. last_success_at: type: - string - 'null' format: date-time description: The last 2xx we received. Null if there has never been one. last_failure_at: type: - string - 'null' format: date-time description: The last delivery that used up its ladder. disabled_at: type: - string - 'null' format: date-time description: When delivery stopped. Null while `active`. disabled_reason: type: - string - 'null' description: >- Why, in prose for a person. Null while `active`. ⚠️ This is the field that answers "why did my endpoint stop receiving events" without a support ticket. created_at: type: string format: date-time description: ISO 8601, UTC. required: - id - url - subscribed_events - brand_id - active - consecutive_failures - last_success_at - last_failure_at - disabled_at - disabled_reason - created_at Error: type: object description: >- The body of every refusal the API returns deliberately, at every status, from every endpoint. Such a response is a failure if and only if it carries an `error` key — that test is correct even for a client that ignores the status line, which is a thing real clients do. An unexpected server failure may not have this shape: treat a 5xx without an `error` key as an `internal_error`, which is retryable. properties: error: type: object description: The refusal. Anything added in future is added INSIDE this object. properties: code: type: string description: >- A stable snake_case identifier that will never change meaning or spelling. This is the contract; branch on it. See `x-error-codes` at the root of this document for retry semantics per code. enum: - unauthenticated - forbidden - insufficient_scope - not_found - invalid_request - rate_limited - internal_error - service_unavailable - idempotency_key_required - idempotency_key_reused - request_in_progress - template_archived - template_unusable - template_has_no_fields - recipient_role_unknown - recipient_role_missing - recipient_role_duplicated - value_subject_unknown - value_address_unknown - value_not_writable - value_not_lockable - locked_answer_missing - locked_answer_unusable - cannot_be_drawn - send_allowance_exhausted - test_key_cannot_send - test_key_cannot_read_documents - test_key_cannot_mint - webhook_url_invalid - webhook_events_invalid - webhook_brand_unknown - webhook_limit_reached - api_key_owner_removed - recipient_not_embedded - recipient_not_yet_turn - recipient_cannot_sign - envelope_not_sent - envelope_not_draft - brand_has_no_embed_origins - recipient_cannot_be_embedded - batch_interrupted - brand_unknown - draft_not_available message: type: string description: >- ⚠️ PROSE FOR A DEVELOPER READING A LOG, and NOT part of the contract. It may be reworded in any release; anything that parses it is broken by design. required: - code - message required: - error securitySchemes: ApiKey: type: http scheme: bearer description: >- `Authorization: Bearer vsk_live_…`. Chosen over a bespoke `X-API-KEY` header because every client, proxy and log-redaction rule already knows this one. What a key may DO is its scopes — see `x-scopes` at the root of this document and `x-required-scope` on each operation. The scope list is not written here because OpenAPI reserves a requirement’s scope array for `oauth2` and `openIdConnect` and requires it to be empty for an `http` scheme. ````