Skip to content
Talk to our solutions team

What are Blocks?

kis.ai offers a range of enterprise-grade services, or “blocks,” which developers can leverage to build robust enterprise applications. These blocks are easily configured with YAML and extended with JavaScript and WebAssembly (WASM).

AI-native infrastructure. Every model call routes through the gateway, which is what makes per-tenant budgets, one audit trail and real cost accounting possible.

  1. AI Gateway: One endpoint for every provider — aliasing, fallback, caching, budgets, prompt registry.
  2. AI Flow: Author flows as DAGs of tasks; harnesses add bounded agentic loops.
  3. AI Bot: Conversational infrastructure with a cost-escalating NLU cascade.
  4. AI Knowledge: Lexical, semantic and hybrid retrieval over your corpus.
  5. AI Memory: Short- and long-term memory with compaction and decay.
  6. AI Evals: Measure quality and gate deploys on it.

The backend an enterprise application needs. The first group is the foundation everything else runs on.

  1. Config: Runtime configuration control plane plus TTL-leased service discovery — config.svc.
  2. Meta: Serves each product’s monorepo to the generic service blocks — meta.svc.
  3. Vault: Secrets, zero-trust mTLS PKI and service-to-service JWTs — vault.svc + vault-agent.
  4. IAM: Authentication methods, authorization rules, tokens and realms — iam.svc.
  5. Data API: REST, GraphQL, gRPC and OData from YAML entity definitions.
  6. Gateway: TLS-terminating multi-tenant edge proxy — gateway.svc.
  7. BFF: One endpoint per screen, stitched from several backends.
  8. Script: Sandboxed multi-language execution of tenant code — script.svc.
  9. Content: Headless content management with typed models and asset stores.
  10. Intake: Hardened anonymous form capture at the edge.
  11. Workflows: Long-running processes that suspend, wait for humans and resume.
  12. Jobs: Scheduled and triggered background work on a fleet of agents.
  13. Rules: Decision logic and spatial document analysis — extraction, matching, validation.
  14. Data Pipes: Scalable ETL and ELT pipelines.
  15. Integrate: Named connections to third-party systems.
  16. Notifications: Omni-channel dispatch with routing, suppression and an outbox.
  17. Audit: Tamper-evident audit trail.
  18. Usage: Telemetry and time decomposition — human versus AI versus idle.

Frontend building blocks. Most libraries currently target Vue; React support is in progress.

  1. Components: Enterprise components — forms, grids, charts.
  2. Theming: Universal theming built on UnoCSS.
  3. Figma: Convert Figma designs into components and themes.
  4. Localization: Multiple languages and locales.
  5. Accessibility: Components aligned to WCAG 2.2.
  6. JS Libraries: Client libraries for the AI and BaaS blocks.

The inner loop — authoring, verifying and shipping.

  1. Spec: The YAML schema system — 128 kinds, validation, linting, scaffolding, generation.
  2. IDE: Language server for platform YAML — diagnostics, completion, go-to-definition.
  3. Testing: Declarative API and load testing with assertions and data-driven cases.
  4. Code Quality: Static analysis and security scanning, with quality tracked over time.

Running it in production.

  1. Automate: Distributed CI/CD engine.
  2. Observability: Per-tenant collection of logs, traces and metrics, and the query side.
  3. Registries: Managed artifact registries with whitelisting and revocation.
  4. Supervisor: Keeps declared services in their desired state on a node.