Book a demo
About OCNORA

What is OCNORA?

Coordinate AI, workflows and software delivery across your business.

Name
OCNORA
Category
workflow automation and AI-agent platform
Company
KXN Technologies
Domain
ocnora.ai

What is OCNORA?

OCNORA is a workflow automation and AI-agent platform built by KXN Technologies (ocnora.ai). Its category line is “Coordinate AI, workflows and software delivery across your business.” It connects 1,075 cataloged app and cloud-service providers, runs AI agents inside token and USD budgets and human approvals, keeps a record of every run, and deploys on OCNORA Cloud, on the customer’s own cloud, or embedded in the customer’s product.

Sources in the claims registry
  • providers (providers; measured; checked 2026-09-22; source: src/data/providers-index.json#providerCount (npm run sync:catalog))
  • agentNode (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • aiBudgetUsage (budget model; unverified; checked 2026-09-28; source: coax-backend AI cost ledger, budget router and circuit breaker (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • humanApprovals (approval model; unverified; checked 2026-09-28; source: coax-backend human.prompt node and router (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • runAudit (audit model; unverified; checked 2026-09-28; source: coax-backend run store and audit log (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • deploymentOptions (deployment options; unverified; checked 2026-09-28; source: coax-backend deployment manifests: Docker Compose file and Helm chart (evidence register C14, C15, C22; E08, E09, E17 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-036)))

Who is OCNORA for?

OCNORA is for IT operations teams automating requests, approvals and incident follow-up; engineering leaders coordinating software delivery and AI agents in validated workflows; data and platform teams moving data between warehouses, queues and cloud services on their own infrastructure; SaaS product teams embedding automation in their own product under their own brand; and security and procurement teams who need to know where secrets are kept, who can act, what AI can spend and what ran. Customers who have approved being named: SyncRivo, UpSkillZone and ATS Eye.

Sources in the claims registry
  • approvals (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • workflowHealth (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • composeAndHelm (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • embedJitSso (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • whiteLabel (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • encryptionBackends (key backends; unverified; checked 2026-09-27; source: coax-backend KMS adapters (local, GCP and AWS implemented; Azure Key Vault and HashiCorp Vault are stubs))
  • customerNames (named customers; unverified; checked 2026-09-28; source: OCNORA_WEBSITE_IMPROVEMENTS sheet, task WEB-000 (customer names approved for publication))

How is OCNORA different?

In OCNORA the controls are part of the engine on every deployment rather than an add-on: every model call is written to a ledger with its tokens and USD cost and checked against monthly budgets per workspace and per provider (per tenant on Embedded); a human.prompt step holds a run until a person answers; Workflow Health validates each workflow on save and publish and a dry run executes the graph with side effects suppressed; and each run keeps its steps with input, output, status and the connection used. The same product can run on the customer’s own cloud under an offline signed licence with no phone-home, or be embedded in another product with JIT SSO and per-tenant budgets. It imports n8n workflow JSON. Comparisons with other platforms, each figure sourced and dated, are at https://ocnora.ai/compare.

Sources in the claims registry
  • aiBudgetUsage (budget model; unverified; checked 2026-09-28; source: coax-backend AI cost ledger, budget router and circuit breaker (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • humanApprovals (approval model; unverified; checked 2026-09-28; source: coax-backend human.prompt node and router (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • workflowHealth (health check model; unverified; checked 2026-09-28; source: coax-backend Workflow Health role, Auto Fix worker and the system-workflow-health template in coax-backend src/mantle/templates/data, which the site does not publish (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • dryRun (test mode; unverified; checked 2026-09-28; source: coax-backend Workflow Health dry run (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • runAudit (audit model; unverified; checked 2026-09-28; source: coax-backend run store and audit log (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))
  • offlineLicence (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • embedJitSso (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • perTenantBudgets (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • n8nImport (n8n node types (approximate); unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))

How can OCNORA be deployed?

OCNORA runs on OCNORA Cloud, on the customer’s own cloud from the Docker Compose file or the Helm chart in the backend repository, or embedded in the customer’s product. The Helm chart runs each process role as its own deployment, including Workflow Health, Auto Fix and the task orchestrator. Self-hosted and embedded installations verify their licence offline and keep workflow data, run history and secrets in the customer’s own cloud account, region and database. Docker Compose and Helm are the only manifests; running the customer’s own cloud is the customer’s job.

Sources in the claims registry
  • deploymentOptions (deployment options; unverified; checked 2026-09-28; source: coax-backend deployment manifests: Docker Compose file and Helm chart (evidence register C14, C15, C22; E08, E09, E17 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-036)))
  • offlineLicence (deployments; unverified; checked 2026-09-27; source: src/app/pricing/page.tsx feature matrix as published at 5405ddc (coax-backend not re-read))
  • dataHandling (processing region; unverified; checked 2026-09-28; source: coax-backend deployment manifests and offline licence check; evidence register C22 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-037))

Which AI model providers does OCNORA support?

OCNORA wires 4 hosted model vendors for AI building and agents: Anthropic, OpenAI, Google and DeepSeek. Self-hosted and embedded installations can also use local models through any OpenAI-compatible inference server the customer hosts (vLLM, Ollama, LM Studio, llama.cpp). Every model call routed through OCNORA, whichever vendor serves it, is written to the cost ledger and checked against the token and USD budgets.

Sources in the claims registry
  • llmVendors (hosted LLM vendors; unverified; checked 2026-09-27; source: coax-backend AI vendor factory)
  • localModels (deployments; implemented; checked 2026-10-20; source: coax-backend ai/providers/factory.py and ai/providers/openai_compatible_adapter.py (the openai_compatible provider))
  • aiBudgetUsage (budget model; unverified; checked 2026-09-28; source: coax-backend AI cost ledger, budget router and circuit breaker (evidence register C12, C13, C16; E07 (OCNORA_WEBSITE_IMPROVEMENTS task WEB-033)))

How many integrations does OCNORA have?

OCNORA counts integrations in three units that are not interchangeable. Its public catalogue has 1,075 providers (the vendors and individual cloud services it connects to, one integration page each) and 7,621 documented node types: 7,378 belong to a provider and 243 are core node types (control flow, data, AI, files, vectors) that need no connection. 84 of the node types are triggers. A provider is not a node type, and neither is an “app” or an “integration” in another vendor’s sense. A service with no typed node is reached through the HTTP node with the customer’s own credential.

Sources in the claims registry
  • providers (providers; measured; checked 2026-09-22; source: src/data/providers-index.json#providerCount (npm run sync:catalog))
  • nodeTypes (node types; measured; checked 2026-09-22; source: src/data/providers-index.json#documentedNodeTypes (npm run sync:catalog))
  • providerNodeTypes (node types; measured; checked 2026-09-22; source: src/data/providers-index.json#nodeCount (npm run sync:catalog))
  • coreNodeTypes (node types; measured; checked 2026-09-22; source: src/data/providers-index.json#coreNodeTypes (npm run sync:catalog))
  • triggers (trigger types; measured; checked 2026-09-22; source: src/data/providers-index.json#triggerCount (npm run sync:catalog))

The same passages are in llms.txt. For detail, see the platform overview, deployment and the integration catalogue.

See a workflow run on your own use case.

Book a technical demo. Bring an n8n file or pick a template, and we build it with you on the canvas while Workflow Health checks it before it ever runs.

Book a technical demoTalk to sales