Mini Apps

Custom Build. Scoped per engagement.

A framework plus orchestrator for small, purpose-built automations: the ones that don’t fit a clean LLM-only solution. Think extracting structured data from messy PDFs, generating tailored reports or custom workflow logic for non-engineering teams. Mini apps live in a central catalog, and your team invokes them from Slack, Zapier, n8n, Tines or plain REST.

The point isn’t to compete with n8n or Glean. It’s the layer underneath: a governed registry and execution layer, so the same automation is callable from any tool without bespoke per-platform glue.

BYOC (bring your own cloud): it deploys to your AWS, GCP, Azure or on-prem environment. Same architecture, different managed services per cloud.

What you get

• A two-service deployment (API + runner) in your cloud account
• Catalog, queueing, artifact store and queryable logs, on your cloud’s managed services
• A connector framework for Slack, Zapier, n8n, Tines and REST
• 1–2 example mini apps that demonstrate the convention
• Auth, versioning, monitoring and error-alerting hooks built in
• Horizontal scaling through queue-decoupled workers

Tools / stack

• Python + FastAPI: API and runner services
• Managed queueing: AWS SQS, GCP Pub/Sub or Azure Service Bus
• Document store: DynamoDB, Firestore or CosmosDB
• Object storage: S3, GCS or Azure Blob
• Analytics: Athena, BigQuery or Azure Data Explorer
• Slack / Zapier / n8n / Tines / REST: native invocation surfaces
• Your existing observability stack: CloudWatch, Stackdriver, Datadog, Grafana, etc.

When it fits

You have small automations scattered across n8n, Zapier, scripts and “that one engineer’s laptop”, and you want a single layer your team can build into without picking sides on the workflow tool. Deployed in your cloud, owned by your team. It’s the same playbook our founder has built in production.