> ## Documentation Index > Fetch the complete documentation index at: https://developers.nuvia.ai/llms.txt > Use this file to discover all available pages before exploring further. # Update knowledge item or media > Updates the name and/or description of a knowledge item or media ## OpenAPI ````yaml /en/openapi.json patch /v1/knowledge/{id} openapi: 3.0.0 info: title: Nuvia API - Company description: API documentation for companies version: '1.0' contact: {} servers: - url: https://api.nuvia.ai security: [] tags: - name: messages x-group: Mensagens - name: contacts x-group: Contatos - name: campaigns x-group: Campanhas - name: tables x-group: Tabelas e Listas - name: agents x-group: Agentes - name: conversations x-group: Conversas - name: inboxes x-group: Caixas de Entrada - name: knowledge x-group: Base de Conhecimento - name: Webhooks x-group: Webhooks - name: campaign-enrollments x-group: Inscrições em Campanha - name: custom-fields x-group: Campos Customizados paths: /v1/knowledge/{id}: patch: tags: - knowledge summary: Update knowledge item or media description: Updates the name and/or description of a knowledge item or media operationId: KnowledgeController_update parameters: - name: x-company-id in: header description: >- Target company identifier. Required only for global API keys (type=global). Ignored for company API keys and human users. required: false schema: type: string - name: id required: true in: path description: Knowledge/media ID schema: type: string responses: '200': description: '' content: application/json: schema: $ref: '#/components/schemas/KnowledgeEntity' security: - JWT-auth: [] components: schemas: KnowledgeEntity: type: object properties: _id: $ref: '#/components/schemas/ObjectId' name: type: string description: type: string type: enum: - KNOWLEDGE - MEDIA type: string s3_key: type: string description: S3 object key. Source of truth for generating the signed URL. file_url: type: string deprecated: true file_type: enum: - PDF - DOCX - AUTCL - IMAGE - VIDEO - AUDIO - DOCUMENT type: string content: type: string status: enum: - PROCESSING - READY - ERROR type: string meta: type: object company: $ref: '#/components/schemas/CompanyEntity' createdAt: format: date-time type: string updatedAt: format: date-time type: string required: - _id - name - type - file_type - company ObjectId: type: object properties: {} CompanyEntity: type: object properties: _id: $ref: '#/components/schemas/ObjectId' name: type: string channels: type: array items: type: string enum: - WAP_AUTO_CLOSER - WEB_V2 - EVOLUTION_API - LINKEDIN - EMAIL agent_flows: type: array items: $ref: '#/components/schemas/AgentFlowEntity' description: type: string description: >- Optional since ISSUE-6330: companies created via self-service are born without a description (legacy ones from before 2025-09-07 also lack the field). The requirement remains in the backoffice's INPUT layers (Zod DTOs for create-company/create-internal-company, and forms) — untouched. domains: description: >- Email domains that resolve to THIS company in self-service signup (ISSUE-6330): 1 domain → at most 1 company (partial unique index below). Only the self-service flow writes; legacy backoffice companies are left WITHOUT the field (no backfill — they go through resolution by inference via users' email). Canonicalized lowercase/trim in the schema (mirrors signup_leads.domain). `default: undefined` so the field is born ABSENT (a default array [] would enter the partial index's semantics and needlessly change backoffice documents). type: array items: type: string site: type: string description: >- [ISSUE-6438] Site entered in the wizard. INFORMATIONAL — it does not take part in domain-based resolution and has no unique index. It exists because of the personal email workspace (SIGNUP_ALLOW_PERSONAL_EMAILS): there the domain can't be derived from the email, so the user types it in. Putting it in `domains` would claim the domain globally (the index is unique) and would let someone with @gmail register "google.com" — hence a separate field, with `domains: []`. general_config: type: object status: enum: - ACTIVE - INACTIVE - DELETED type: string agent_builder_mcp_enabled: type: boolean description: >- Enables the agent builder tools in the client's `/mcp` for THIS company. Default `true`: the builder is GA — every company has it, unless explicit opt-OUT (`false`). It works as a per-company EXCLUSION switch (e.g., turning off a client that made a mess), not an inclusion one. The AUTHORING gate remains the role (COMPANY_ADMIN); a regular user can only read/simulate. Companies created before GA were enabled via backfill (scripts/backfill-agent-builder-mcp-flag.ts). client_mcp_servers_enabled: type: boolean description: >- Enables the client's MCP servers feature (SDD "Client MCP in agent tools", §8 rollout) for THIS company. Absent/false = flag OFF: the `/mcp-servers` routes respond 403 (except the OAuth callback, which is public) and MCP tools already materialized become INERT at runtime (they disappear from the agent's array, without deleting anything from the database). Deliberate default `false`: it's a gradual rollout, not GA like `agent_builder_mcp_enabled`. auto_join_enabled: type: boolean description: >- [ISSUE-6432] Automatic domain-based entry. Absent/false = OFF (deliberate default: turning it on is an admin decision, not the initial state). When on, anyone who signs up with an email from this company's domain joins directly as a member, instead of generating an access request. Affects ONLY this company: a domain shared by multiple companies still lets the person choose, and among the chosen ones only the one that turned it on grants direct entry. subscription: type: object description: >- Subscription/trial (ISSUE-6334). Default `{status:'active'}` — a company created by the backoffice is born ACTIVE (product decision); the self-service flow (ProvisionCompanyByDomainService) overwrites it with `trialing` + dates. A legacy doc without the field = active (see resolveSubscriptionStatus + the backfill script). onboarding: type: object description: >- [ISSUE-6438] Onboarding wizard state. `default: undefined` so the field is born ABSENT (same technique as `domains`): a backoffice company does not get the subdocument and resolves as completed in `resolveOnboardingStatus`. attribution: type: object description: >- Campaign that brought in this company (`utm_*` from whoever created it). ── Why here, if the UTM already lives in `signup_leads` ── There the key is the EMAIL, and the link only goes as far as the user. Cross-referencing campaign with trial, activation, or revenue required a manual join by the first user's email — which breaks as soon as the person changes email or leaves. Copied at CREATION time, and only then. There's no backfill: a company created before this is left without the field, and absent means "we don't know", not "organic". `default: undefined` so it's born ABSENT (same technique as `onboarding`): a backoffice company doesn't go through the self-service flow and shouldn't get an empty subdocument. required: - _id - name - channels - agent_flows - description - status AgentFlowEntity: type: object properties: _id: $ref: '#/components/schemas/ObjectId' name: type: string url: type: string specialist_url: type: string type: enum: - N8N - API - INTERNAL type: string required: - _id - name - url - specialist_url - type securitySchemes: JWT-auth: scheme: bearer bearerFormat: JWT type: http description: Token JWT de autenticação ````