Jentic Permissions API
Part of the **Admin / Audit** bounded context — the catalogue of grantable permission strings. Permissions are namespaced `resource:action` strings (`users:write`, `capabilities:execute`, `credentials:read`, …) gating every other surface in the API. The set is platform-defined and small; `GET /permissions` returns it in full, with each entry describing what it does, what it implies (the static implication map; e.g. `capabilities:execute` implies `apis:read` and `executions:read`), and whether the calling user is authorised to grant it. One reserved superpower short-circuits individual checks in code: `org:admin` (full deployment-wide access, granted via direct DB action). It is not enumerated to non-holders by `GET /permissions`, and is rejected by `PUT /users/{user_id}/permissions` from any caller who doesn't already hold it. The same vocabulary is used for `User.permissions` — coarse JWT-embedded scopes — so the catalogue below covers user assignment. Toolkits have no separate scope tier of their own: a toolkit API key is minted with the fixed broker-execute scope, and the per-binding fine-grained `PermissionRule[]` (the inner PBAC tier) lives separately under the `Toolkit Permissions` tag.
GET
/permissions
List Permissions
#
Every API here is available over the APIs.io API and to AI agents over MCP.
MCP server
One button, every client — Claude, Cursor, VS Code and the rest.
https://apis.io/mcp
Tools for apis
7 MCP tools reach this
find_apisBrowse and filter every API in the catalog.
get_api_artifactsOne API's artifacts, grouped by type.
get_openapiThe primary OpenAPI for this API.
find_similar_apisAPIs that look like this one.
apis_io_searchSTART HERE — APIs, providers and tags for one query, each with its total.
resolveTurn a domain, URL or GitHub org into the provider it belongs to.
find_cohortsEvery scored population of providers in the catalog.
All 92 tools →
Call it yourself
curl for this page
This API
curl "https://apis.io/api/v1/apis/jentic:jentic-permissions-api"
All apis
curl "https://apis.io/api/v1/apis?limit=25"
Calls without X-API-Key need no key. Calls with it need a paid plan — get a key.
Get an API key
Free tier, no form to fill in. Signing in shares your email address with us — we
store it to create your key and to recognise you if you sign in with another
provider. See our Privacy Policy and
Terms.
A second provider on the same verified email joins the account you already have.
openapi: 3.2.0
info:
contact:
email: support@jentic.com
name: Jentic Platform Team
url: https://jentic.com/
description: '## Overview ##
The **Jentic Control Plane API** is the unified HTTP surface of the
Jentic platform''s control plane — every administrative and
metadata-facing endpoint outside the Broker data plane. The control plane is the administrative counterpart to the
Broker data plane: where the Broker executes
upstream API calls on behalf of agents, the control plane is where
humans, operators, and agents configure *what* the Broker is
allowed to do, *for whom*, *with which credentials*, *against
which registered APIs*, and *with what telemetry retained*. ## Components ##
High-level components for the control plane API. - **Core / Access** — credentials, toolkits,
and PBAC grants and runtime params. Owns *who can call what, with which secret*. - **Registry** — registered APIs, immutable
revisions, operations, search, lookup. Owns *what APIs are
registered for this deployment and what they look like at any
given revision*. - **Admin / Audit** — users, permissions,
execution telemetry, async-job tracker, and dashboard event
stream. Owns *who can sign in with what permissions* together
with *what actually happened* — the human roster, the
append-only records of brokered executions, the unified
`Job` resource that both Registry imports and Broker
async-execution calls feed into, and a curated `/events`
surface for dashboards and operators. ## Authentication ##
Today the platform ships a local username + password identity
provider for human users, alongside agent identity (Dynamic Client
Registration + RFC 7523 JWT-bearer assertions) and service-account
client credentials — every authenticated operation expects
`Authorization: Bearer ` (`BearerAuth`, an opaque `at_`
token) except `GET /health`, `POST /auth/login`,
`POST /users:create-admin`, and `POST /users:redeem-invite`. Human tokens are issued by `POST /auth/login` with a fixed 1-hour
TTL and can be re-minted before expiry via `POST /auth/refresh`
(sliding session, bounded by an absolute window —
`admin.auth.session_ttl_seconds`, 12 hours by default); agents and
service accounts obtain tokens from `POST /oauth/token` (see
`BearerAuth`). The `permissions` claim on a token is a snapshot at
issue time; permission changes take effect at the next re-issue —
the next refresh or re-login (≤ 1 hour with the default TTL). The platform ships **no default credentials**. On a fresh
install the users table is empty, so `GET /health` returns
`setup_required: true` with `next_step: create_admin`. The
operator creates the first administrator with `POST
/users:create-admin {email, password}` — a one-time,
unauthenticated, self-closing endpoint: it succeeds only while
the users table is empty and returns `410 setup_already_complete`
once any user exists (a single-row setup sentinel is the
concurrency backstop, so two racing callers — even with different
emails — cannot both bootstrap). It grants `org:admin`, sets the operator''s chosen
password (so `must_change_password` is **false** — no forced
rotation), and returns a ready-to-use `LoginResponse`
(auto-login). After the first user exists `setup_required` flips
to `false`. See the **Users** tag for the full onboarding
walkthrough. External SSO providers (Okta, AWS Cognito, Auth0, Azure AD)
plug in via the `User.auth_provider: external` hook plus an
`external_subject_id` mapping; no concrete provider ships in
this revision. Agent identity (separate from human users) uses
Dynamic Client Registration (`POST /register`) with the agent''s
JWKS, then RFC 7523 JWT-bearer assertions exchanged at
`POST /oauth/token` for an opaque access token. ## Cross-cutting conventions
- Every API reference uses the loose
`APIReference(vendor, name, version)` data structure. There is no
foreign key between Core, Registry, and Admin / Audit — the
tuple is the only join. - **Error shape.** Every error response is RFC 9457 Problem
Details (`application/problem+json`) using the shared
`ProblemDetails` schema from
jentic/api-problem-details.'
license:
identifier: Apache-2.0
name: Apache-2.0
summary: HTTP surface of the Jentic platform's control plane — Core / Access (credentials, toolkits), Registry (registered APIs, operations, search), and Admin / Audit (execution telemetry, async jobs).
title: Jentic Control Plane Permissions API
version: 0.39.0
servers:
- description: Same-origin (relative)
url: /
security:
- BearerAuth: []
tags:
- description: 'Part of the **Admin / Audit** bounded context — the catalogue of grantable permission strings. Permissions are namespaced `resource:action` strings (`users:write`, `capabilities:execute`, `credentials:read`, …) gating every other surface in the API. The set is platform-defined and small; `GET /permissions` returns it in full, with each entry describing what it does, what it implies (the static implication map; e.g. `capabilities:execute` implies `apis:read` and `executions:read`), and whether the calling user is authorised to grant it.
One reserved superpower short-circuits individual checks in code: `org:admin` (full deployment-wide access, granted via direct DB action). It is not enumerated to non-holders by `GET /permissions`, and is rejected by `PUT /users/{user_id}/permissions` from any caller who doesn''t already hold it.
The same vocabulary is used for `User.permissions` — coarse JWT-embedded scopes — so the catalogue below covers user assignment. Toolkits have no separate scope tier of their own: a toolkit API key is minted with the fixed broker-execute scope, and the per-binding fine-grained `PermissionRule[]` (the inner PBAC tier) lives separately under the `Toolkit Permissions` tag.'
name: Permissions
paths:
/permissions:
get:
description: List the permission catalogue visible to the caller.
operationId: listPermissions
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/PermissionListResponse'
description: Successful Response
'400':
content:
application/problem+json:
example:
detail: The request was malformed or failed a precondition.
instance: /example/path
status: 400
title: Bad Request
type: bad_request
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Bad Request
'401':
content:
application/problem+json:
example:
detail: Authentication is required or the bearer token is invalid.
instance: /example/path
status: 401
title: Unauthorized
type: unauthorized
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Unauthorized
'403':
content:
application/problem+json:
example:
detail: The authenticated principal lacks the required permission.
instance: /example/path
status: 403
title: Forbidden
type: forbidden
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Forbidden
'422':
content:
application/problem+json:
example:
detail: Request validation failed; see errors[] for details.
errors:
- detail: Field 'name' must not be blank.
pointer: '#/name'
instance: /example/path
status: 422
title: Unprocessable Entity
type: validation_error
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Unprocessable Entity
'500':
content:
application/problem+json:
example:
detail: An unexpected error occurred.
instance: /example/path
status: 500
title: Internal Server Error
type: server_error
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Internal Server Error
'503':
content:
application/problem+json:
example:
detail: The database is temporarily unavailable; please retry.
instance: /example/path
status: 503
title: Service Unavailable
type: database_unavailable
schema:
$ref: '#/components/schemas/ProblemDetail'
description: Service Unavailable
security:
- BearerAuth: []
summary: List Permissions
tags:
- Permissions
components:
schemas:
ProblemDetail:
description: 'RFC 9457 Problem Details for HTTP APIs.
This is the standard error response format for all Jentic APIs.
Content-Type: application/problem+json
See: https://www.rfc-editor.org/rfc/rfc9457.html'
examples:
- detail: The request body is missing one or more required fields.
errors:
- detail: Field 'name' is required.
pointer: '#/name'
instance: /v2/capability-sets
status: 400
title: Bad Request
type: about:blank
properties:
code:
anyOf:
- maxLength: 50
type: string
- type: 'null'
description: An optional provider-specific code for internal error taxonomy and observability correlation.
examples:
- JENTIC-4001
title: Code
detail:
description: A human-readable explanation specific to this occurrence of the problem. MUST be present. Provide actionable information where possible.
examples:
- The request body is missing required field 'name'.
maxLength: 4096
title: Detail
type: string
errors:
anyOf:
- items:
$ref: '#/components/schemas/ErrorItem'
maxItems: 1000
type: array
- type: 'null'
description: An array of granular error details. Use when multiple validation errors or field-level problems need to be surfaced in a single response.
title: Errors
instance:
anyOf:
- maxLength: 1024
type: string
- type: 'null'
description: A URI reference identifying the specific occurrence of the problem. Typically the request path.
examples:
- /v2/capability-sets
title: Instance
status:
anyOf:
- maximum: 599.0
minimum: 100.0
type: integer
- type: 'null'
description: The HTTP status code for this occurrence of the problem.
examples:
- 400
title: Status
title:
anyOf:
- maxLength: 1024
type: string
- type: 'null'
description: A short, human-readable summary of the problem type. Should not change between occurrences except for localisation purposes.
examples:
- Bad Request
title: Title
type:
default: about:blank
description: A URI reference identifying the problem type. When set to 'about:blank', the title SHOULD be the standard HTTP status phrase. Use an IANA-registered type URI where one applies.
examples:
- about:blank
maxLength: 1024
title: Type
type: string
required:
- detail
title: ProblemDetail
type: object
ErrorItem:
description: 'A granular error detail entry within the errors[] array of a ProblemDetails response.
At least one of pointer, parameter, or header SHOULD be present to identify the error source.'
properties:
code:
anyOf:
- maxLength: 50
type: string
- type: 'null'
description: An optional provider-specific code identifying this error in internal taxonomy or documentation.
examples:
- JENTIC-V-001
title: Code
detail:
description: A human-readable explanation of this specific error. Be precise — name the field, parameter, or header involved.
examples:
- Field 'name' must not be blank.
maxLength: 4096
title: Detail
type: string
header:
anyOf:
- maxLength: 1024
type: string
- type: 'null'
description: The name of the request header that is the source of this error.
examples:
- Authorization
title: Header
parameter:
anyOf:
- maxLength: 1024
type: string
- type: 'null'
description: The name of the query or path parameter that is the source of this error.
examples:
- limit
title: Parameter
pointer:
anyOf:
- maxLength: 1024
type: string
- type: 'null'
description: A JSON Pointer (RFC 6901) to the specific request body property that is the source of this error.
examples:
- '#/name'
title: Pointer
required:
- detail
title: ErrorItem
type: object
PermissionListResponse:
description: List of available permissions.
properties:
data:
items:
$ref: '#/components/schemas/PermissionResponse'
title: Data
type: array
required:
- data
title: PermissionListResponse
type: object
PermissionResponse:
description: A single permission entry from the catalogue.
properties:
description:
title: Description
type: string
grantable_by_caller:
title: Grantable By Caller
type: boolean
implies:
items:
type: string
title: Implies
type: array
name:
title: Name
type: string
required:
- name
- description
- implies
- grantable_by_caller
title: PermissionResponse
type: object
securitySchemes:
BearerAuth:
description: 'Opaque bearer access token (prefixed `at_`), sent as `Authorization: Bearer <token>`. It is **not** a JWT — it is a random string validated server-side by lookup, so it cannot be decoded by the client.
Obtain one from `POST /oauth/token` using the grant for your actor type:
- **Agents** — `grant_type=urn:ietf:params:oauth:grant-type:jwt-bearer` with a JWT assertion signed by the key registered via `POST /register` (the JWT is the *assertion*, not the resulting access token).
- **Service accounts** — `grant_type=client_credentials` with `client_id` + `client_secret`.
- **Users** — `grant_type=authorization_code` (interactive) or `grant_type=password`; refresh either with `grant_type=refresh_token`.
Service accounts can also mint short-lived, scope-narrowed task tokens for agents via `POST /oauth/mint`.
Per-endpoint scope and actor-type requirements are not modelled in this document (OpenAPI cannot faithfully express the OR-of-scopes / `org:admin` bypass / service-layer enforcement); see `GET /reference/endpoints.json` for the authoritative authorization reference.'
scheme: bearer
type: http
x-tagGroups:
- name: Core / Access
tags:
- Credentials
- Toolkits
- Toolkit Keys
- Toolkit Credentials
- Toolkit Permissions
- Access Requests
- name: Registry
tags:
- APIs
- API Operations
- API Spec
- Catalog
- Search
- Inspect
- Lookup
- Notes
- Overlays
- name: Admin / Audit
tags:
- Users
- Actors
- Permissions
- Executions
- Jobs
- Events
- Audit
- Monitoring
- Configuration
- OAuth Clients
- name: Platform Actors
tags:
- Identity
- Agents
- Service Accounts
- OAuth
- Agent Registration
- Discovery
- name: Operations
tags:
- System
- MCP