Independent software company · Nairobi

Built to keep running.

We design, build, and operate infrastructure and autonomous systems where reliability is the product: delivery guarantees you can audit, agents you can supervise, and software a stranger could run at 3 a.m.

135+ verified integrationsExactly-once deliveryEvery attempt traced
ops feed · silven dynamics
01 · Products

Two systems, one philosophy.

Nothing is trusted until it is logged, retried, and reviewable.

MTR-01In service

Mittr Webhook infrastructure you don't have to build

Send the event. Mittr does the rest. It persists every event before delivery, retries until it lands, signs each attempt, and accounts for every one. Free tier, no card, live at mittr.io.

Delivery
Transactional outbox · exactly-once
Scheduling
10 priority levels · fair per tenant
Security
HMAC · mTLS · SSRF guards
Integrations
135+ inbound · S3/GCS sinks
SYS-01In development

SYS_01 Autonomous accounts receivable

An agent that chases invoices so small businesses don't have to. It drafts, evaluates, and queues customer outreach, with autonomy that is earned, logged, and reviewable at every step.

Pipeline
Draft → grade → approve → send
Policy
Hard checks, set per client
Oversight
Human approval queue
Audit
Append-only event log
02 · Capabilities

What we build for ourselves is what we build with partners.

C-01

Infrastructure engineering

Event-driven backends in Go: webhook delivery, queues, retries, backpressure, and observability. The unglamorous plumbing that decides whether a product can be trusted.

C-02

Autonomous systems

LLM agents for real business processes, with evaluation frameworks, model-based grading, and human approval designed in from the first commit rather than bolted on after.

C-03

Product engineering

Full-stack, multi-tenant SaaS with Laravel, Vue, and PostgreSQL, from schema design to interface. Shipped fast without giving up durability.

BOM · Standard parts
  • Go
  • PostgreSQL
  • Redis
  • Laravel
  • Vue 3
  • TypeScript
  • GraphQL
  • Docker
  • Kubernetes
  • OpenTelemetry
  • Claude API
  • Playwright
03 · Principles

Three rules, applied to everything that ships.

P-01

Durable by design

Guarantees live in the database, not in hope. Transactional outboxes, idempotency keys, and append-only audit trails mean failure is planned for before launch, not patched after.

P-02

Humans stay in the loop

Autonomous systems earn their autonomy. Every agent action is graded, recorded, and reviewable, and a person can take over at any step.

P-03

Boring on purpose

Postgres, queues, and plain code. We choose the technology that will still make sense to whoever operates the system years from now.

Working on something
that has to stay up?

Talk to us about Mittr, early access to SYS_01, or building something together.