client.ts
Source: sdks/typescript/packages/sdk/src/client.ts. Back to the TypeScript SDK reference index.
note
This page is generated by tools/docs/gen_sdk_reference.py from the SDK sources and the OpenAPI text folders. Do not edit it by hand: change the source and regenerate.
Type alias EnterpriseAIOptions
Options of the EnterpriseAI constructor, the same as ClientOptions of the HTTP client.
Definition: ClientOptions.
Class EnterpriseAI
Client of the Enterprise AI Control Plane API v1. One sub-client per API resource.
Attributes:
| Name | Type | Required | Description |
|---|---|---|---|
a2a | A.A2AResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
adapters | L.AdaptersResource | yes | LLM and Model Gateway. Model providers, the model registry, routing rules and the Model Gateway: the only path agents and applications use to call models. |
agents | A.AgentsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
apiKeys | I.ApiKeysResource | yes | Identity. Who calls the API and with which rights. This tag covers the caller identity (whoami), API keys for scripts and services, service identities (SPIFFE workloads) that a user may temporarily act as, and the directory group mappings that turn Keycloak groups into project memberships and quotas. |
applications | Tr.ApplicationsResource | yes | Governed applications and third party agents. |
approvals | I.ApprovalsResource | yes | Approvals. Human in the loop: approval requests for sensitive actions, with separation of duties, quorum and expiry. Agent runs that reach a tool needing approval, promotions of models, agents, tools, skills, MCP servers or plugins, break glass grants and mandates all create approval requests here. |
audit | M.AuditResource | yes | Audit events (/v1/audit): listing, hash chain check, WORM export and correlations. |
authz | I.AuthzResource | yes | OpenFGA checks and tuples and explained OPA decisions; an answer without an explicit decision throws AuthzUnconfirmedError. |
breakGlass | Tr.BreakGlassResource | yes | Break-glass: time boxed emergency access, audited and revocable. |
businessUnits | I.BusinessUnitsResource | yes | Tenant hierarchy (ADR-0156): business units of an organization, optionally nested. |
catalog | I.CatalogResource | yes | System. Technical endpoints of the Control Plane: liveness and readiness probes, description of the instance, identity of the caller, the catalog of platform bricks and the product roadmap. |
channels | T.ChannelsResource | yes | Channels. Channel Gateway: the entry points where end users talk to agents (web, mobile, email, Slack, Teams, WhatsApp, SMS, voice, API, and the Navin channels Discord, Telegram, Matrix, Mattermost and Signal). An inbound message becomes a session message and starts a run of the channel agent, within size, rate and session limits and with DLP. |
checkpoints | A.CheckpointsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
compliance | D.ComplianceResource | yes | Navin Enterprise Compliance & Audit Agent instances: activation and lifecycle owned by the Control Plane, the same model as the RAG Agent instances (ADR-0177 section 8). |
connectors | T.ConnectorsResource | yes | Connectors. Connector Hub: registry of the connections to external systems (SaaS applications, APIs, data sources), their credentials kept server side, synchronisation runs, the Credential Broker that hands short lived leases instead of secrets, and the Webhook Gateway that receives signed provider events. |
context | L.ContextResource | yes | Context Manager plan and signed execution context tokens (/v1/context). |
contracts | I.ContractsResource | yes | Contracts and manifests. Published JSON Schemas (contracts) of the platform documents, and declarative manifests that create or update objects from a document, in the spirit of infrastructure as code. |
credentials | T.CredentialsResource | yes | Connectors. Connector Hub: registry of the connections to external systems (SaaS applications, APIs, data sources), their credentials kept server side, synchronisation runs, the Credential Broker that hands short lived leases instead of secrets, and the Webhook Gateway that receives signed provider events. |
delegations | Tr.DelegationsResource | yes | Delegation chains: a principal hands a restricted subset of its rights to an agent. |
dlp | D.DlpResource | yes | Privacy and data protection. Data protection services of the tenant: classification, DLP (data loss prevention), tokenization and envelope encryption. |
encryption | D.EncryptionResource | yes | Privacy and data protection. Data protection services of the tenant: classification, DLP (data loss prevention), tokenization and envelope encryption. |
evaluations | M.EvaluationsResource | yes | Evaluation and red team. Quality of the AI systems of the tenant: evaluation suites, evaluation runs, quality gates before promotion, A/B comparisons against a baseline and red team runs. |
events | A.EventsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
evidence | P.EvidenceResource | yes | Evidence Service: seals, verification, legal holds, retention and run configuration evidence. |
fileSecurity | T.FileSecurityResource | yes | /v1/file-security: engine status, tenant policy, explicit scans, rescans, sample purge and stats. |
finops | F.FinOpsResource | yes | FinOps. Cost management of AI usage: price cards, usage events, estimates, quotas and reservations, GPU cost allocation, billing reconciliation and alerts. |
gateway | L.GatewayResource | yes | Model gateway: the only path agents use to call models (OpenAI-compatible bodies plus eai). |
governance | I.GovernanceResource | yes | Governance. Governance registry: one aggregated view of the AI assets of the tenant, to know what exists, who owns it and how risky it is. |
heartbeats | A.HeartbeatsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
http | HttpClient | yes | HTTP transport shared by every sub-client. |
incidents | I.IncidentsResource | yes | Incidents. Security and operations incidents of the tenant: severity, status, timeline, containment actions, Alertmanager alerts and forwarding to the SIEM. |
isolation | L.IsolationResource | yes | LLM isolation profile of the tenant or of a project (dedicated deployments, regions, providers). |
knowledgeBases | D.KnowledgeBasesResource | yes | Knowledge bases. Knowledge bases for retrieval augmented generation (RAG): documents are scanned, extracted, screened for personal data and poisoning, chunked, embedded through the Model Gateway and searched with ACL and clearance filters. |
license | Tr.LicenseResource | yes | License entitlement in force for the installation (contract license-entitlement). |
lineage | D.LineageResource | yes | Lineage. Data lineage of the tenant, fed with OpenLineage events: which jobs read which datasets and produce which others. |
litellm | L.LiteLLMResource | yes | LLM and Model Gateway. Model providers, the model registry, routing rules and the Model Gateway: the only path agents and applications use to call models. |
llm | L.LlmResource | yes | Single llm interface: generate, stream, embed and vision, all through the Model Gateway checks. |
llmManagement | LlmManagementResource | yes | LLM projects, quotas and LiteLLM keys (/v1/llm/projects, roles tenant_admin or platform_admin). createKey returns the clear LiteLLM key once in key: the platform never stores it. |
mandates | Tr.MandatesResource | yes | Mandates: standing authority of an autonomous agent, granted through an approval. |
manifests | I.ManifestsResource | yes | Contracts and manifests. Published JSON Schemas (contracts) of the platform documents, and declarative manifests that create or update objects from a document, in the spirit of infrastructure as code. |
marketplace | P.MarketplaceResource | yes | Signed packages, publishers and governed installations. |
mcp | T.McpResource | yes | MCP gateway. Registry of the upstream MCP servers (Model Context Protocol) and the MCP Gateway, a JSON-RPC 2.0 proxy that applies the platform controls to every call: project isolation, allowed callers, allow lists of tools and resources, scopes, DLP, policy, quota, prompt injection screening, metering and audit. |
mcpRuntime | T.McpRuntimeResource | yes | MCP Runtime: isolated MCP instances of the caller in a project, on the ToolHive Kubernetes Operator (ADR-0167). |
me | I.MeResource | yes | Signed in caller: /v1/me, whoami, local account login, password change and logout. |
media | M.MediaResource | yes | Multimodal. Media and multimodal processing: encrypted media storage and jobs for vision, image generation and edition, video processing and generation, speech to text, diarization, text to speech and multimodal DLP. |
members | I.MembersResource | yes | Members of the tenant, quota profiles and directory group mappings. |
memory | D.MemoryResource | yes | Memory. Memory Manager: long lived memory items of agents and users, following the Memory Contract v1, with ACL, lifetime, legal hold, versioning and hybrid search. |
meta | I.MetaResource | yes | System. Technical endpoints of the Control Plane: liveness and readiness probes, description of the instance, identity of the caller, the catalog of platform bricks and the product roadmap. |
models | L.ModelsResource | yes | LLM and Model Gateway. Model providers, the model registry, routing rules and the Model Gateway: the only path agents and applications use to call models. |
multimodal | M.MultimodalResource | yes | Multimodal. Media and multimodal processing: encrypted media storage and jobs for vision, image generation and edition, video processing and generation, speech to text, diarization, text to speech and multimodal DLP. |
navinLibrary | T.NavinLibraryResource | yes | Navin library vendored with the platform, activated per project, and the rules of a project. |
operations | I.OperationsResource | yes | Long operations (architecture B6.3): follow the Operation-Id returned by a long creation. |
organizations | I.OrganizationsResource | yes | Tenant hierarchy (ADR-0156): organizations of the tenant. PATCH with expected_version is optimistic (409 version_conflict). |
platformManager | P.PlatformManagerResource | yes | Platform Manager space of the console (ADR-0128), platform_admin only. The Control Plane relays every call to the Navin Platform Manager on behalf of the caller; a Manager refusal comes back with its status and a platform_manager_<code> error. |
plugins | T.PluginsResource | yes | Plugins. Registry and runtime lifecycle of plugins: signed OCI images run by the platform in a sandbox (gVisor, Kata or Wasm runtime class), with a Kubernetes Deployment and a NetworkPolicy rendered from the manifest. |
policies | I.PoliciesResource | yes | Policies. Attribute based policies (ABAC) that decide every sensitive action of the platform, with the Policy Contract v1, the decision log, simulation, revisions and signed OPA bundles. |
privacy | D.PrivacyResource | yes | Privacy and data protection. Data protection services of the tenant: classification, DLP (data loss prevention), tokenization and envelope encryption. |
projects | I.ProjectsResource | yes | Tenancy. Tenants, their hierarchy, projects, environments and members. |
prompts | D.PromptsResource | yes | Prompts. Prompt registry: versioned prompt templates with typed variables and validated rendering. |
providers | L.ProvidersResource | yes | LLM and Model Gateway. Model providers, the model registry, routing rules and the Model Gateway: the only path agents and applications use to call models. |
quotas | F.QuotasResource | yes | FinOps. Cost management of AI usage: price cards, usage events, estimates, quotas and reservations, GPU cost allocation, billing reconciliation and alerts. |
ragAgent | D.RagAgentResource | yes | Navin Enterprise RAG Agent instances: activation and lifecycle owned by the Control Plane. |
realtime | M.RealtimeResource | yes | Realtime voice. Realtime voice sessions: a LiveKit room token and a WebSocket event protocol to stream audio to an agent, with participant consent, turns, DLP filtered transcripts, optional recordings and metering. |
redteam | M.RedteamResource | yes | Evaluation and red team. Quality of the AI systems of the tenant: evaluation suites, evaluation runs, quality gates before promotion, A/B comparisons against a baseline and red team runs. |
retention | D.RetentionResource | yes | Retention. How long the data of the tenant is kept, how it is purged, the erasure of a data subject (GDPR article 17) and legal holds. |
risk | I.RiskResource | yes | Risk and compliance. Risk management of the AI systems of the tenant: EU AI Act assessments, controls mapped to ISO/IEC 42001, NIST AI RMF, GDPR and the EU AI Act, evidence and time limited exceptions. |
roadmap | I.RoadmapResource | yes | System. Technical endpoints of the Control Plane: liveness and readiness probes, description of the instance, identity of the caller, the catalog of platform bricks and the product roadmap. |
routes | L.RoutesResource | yes | LLM and Model Gateway. Model providers, the model registry, routing rules and the Model Gateway: the only path agents and applications use to call models. |
runs | A.RunsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
sandbox | A.SandboxResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
schedules | A.SchedulesResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
secrets | I.SecretsResource | yes | Secrets. Write only store of the tenant secrets: provider keys, connector passwords, tokens of MCP servers. Values are encrypted (in OpenBao when configured, otherwise with AES-GCM in the database) and are never returned by the API. Other objects refer to a secret by its path; the platform reads the value server side only when it needs it. |
security | I.SecurityResource | yes | Security posture and settings, incident report and run containment. |
serviceIdentities | I.ServiceIdentitiesResource | yes | Identity. Who calls the API and with which rights. This tag covers the caller identity (whoami), API keys for scripts and services, service identities (SPIFFE workloads) that a user may temporarily act as, and the directory group mappings that turn Keycloak groups into project memberships and quotas. |
sessions | A.SessionsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
skills | T.SkillsResource | yes | Skills. Skill Registry and Skill Loader. A skill is a signed, versioned set of instructions (inline, or an artifact in Git or OCI) with the permissions it needs; agents load the skills granted in their manifest. |
soc | S.SocResource | yes | AI-SOC: detections, reaction orders, kill switches verified by receipts and four eyes recovery. |
storageProfiles | P.StorageProfilesResource | yes | Storage. Storage profiles: the client buckets and containers the platform writes to (audit exports, evidence, documents, media, exports, backups), with their credential reference and their capabilities discovered by real calls. |
subagents | A.SubagentsResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
telemetry | M.TelemetryResource | yes | Telemetry. Traces, request journal, live stream and service level objectives (SLO) of the platform. |
telephony | TelephonyResource | yes | Telephony of the tenant: settings, PBX pools, trunks, numbers, extensions, routing, IVR, queues, agents, calls, recordings, voicemail, fraud and infrastructure. |
tenants | I.TenantsResource | yes | Tenancy. Tenants, their hierarchy, projects, environments and members. |
tokenization | D.TokenizationResource | yes | Privacy and data protection. Data protection services of the tenant: classification, DLP (data loss prevention), tokenization and envelope encryption. |
tools | T.ToolsResource | yes | Tool registry and Tool Gateway invocation (/v1/tools). |
triggers | A.TriggersResource | yes | Agents. Registry of the tenant's agents and the server side execution of their runs: steps, budgets, checkpoints, heartbeats, sub-agents, agent to agent (A2A) messages, conversation sessions, cron schedules, event triggers and sandbox executions. |
webhooks | T.WebhooksResource | yes | Connectors. Connector Hub: registry of the connections to external systems (SaaS applications, APIs, data sources), their credentials kept server side, synchronisation runs, the Credential Broker that hands short lived leases instead of secrets, and the Webhook Gateway that receives signed provider events. |
Constructor EnterpriseAI
new EnterpriseAI(options: EnterpriseAIOptions)
Creates the HttpClient from the options and one sub-client per API resource on top of it.
| Name | Type | Default | Required | Description |
|---|---|---|---|---|
options | EnterpriseAIOptions | yes | Per call options (idempotency key, traceparent, headers, abort signal). |