OpsControl brings together secure vault-based access management, real-time service monitoring, and an AI assistant that understands your infrastructure — all in one platform.
Every team above 5 servers hits the same wall. The tools don't talk to each other. The AI doesn't know your infra. And the on-call engineer still wakes up at 3am.
Dozens of servers, SSH keys, DB credentials — scattered across Keepass, shared docs, and someone's memory.
A port closes, a service crashes. Nobody knows until a user reports it — or worse, a customer does.
General-purpose assistants can explain commands, but they should not hold root access or execute changes. They need controlled context, not infrastructure keys.
Who accessed what server, when, with whose credentials? "I think it was Alexey" is not a security posture.
Each module solves a distinct problem. Together, they cover the full ops loop — from access to analysis.
Vault-backed access and secrets layer with infrastructure monitoring. Centralize SSH workflows without giving AI privileged infrastructure control.
AI ops layer that classifies intent, routes to the right model, and answers from your own docs — via voice or text in Telegram.
The full loop: you speak, the platform classifies, routes, checks infrastructure, and answers — all before you finish your coffee.
Send a voice message or text query to your Telegram bot — in any language, any format
Whisper transcribes audio, SetFit classifies complexity in ~20ms, routes to the right model
If needed, InfraConnect scans relevant servers and ports — real data, not guesses
Response lands in Telegram with context, server status, and suggested action
The classifier separates "what time is it?" from "design a fault-tolerant database cluster" — so simple queries don't cost you GPT-4 tokens.
No reinventing the wheel. Every component is production-tested open source — assembled into a coherent platform.
Secrets engine, dynamic credentials, AppRole auth
OpenAI's ASR model — GPU-accelerated, multilingual
Few-shot sentence classifier — trained on custom ops data
Workflow orchestration — routes between models and services
Backend API layer — typed, async, production-ready
Job queue for async audio processing, retry logic
Server registry, port config, credential metadata
Full stack in one command — self-hosted, on-premise
Small and mid-size engineering teams who run real infrastructure but can't afford enterprise tooling — or the overhead that comes with it.
Stop memorizing 50 passwords. One master key, auto-rotating credentials, SSH in 3 seconds.
Ask "why is staging slow?" by voice at 2am. Get an actual answer with port data, not a generic tutorial.
Full audit log of who accessed what. Credential rotation on demand. Compliance-ready from day one.
Workflow-AI handles intent — InfraConnect handles action. They share context through the OpsControl core.
Transparent by default. Here's the honest state of the project.
HashiCorp Vault integration, credential storage, SSH rotation, port monitoring, audit logging, brute-force protection.
ShippedWhisper transcription, SetFit classifier, Agent Router, n8n orchestration, Telegram voice interface.
In ProgressInfraConnect + Workflow-AI connected: assistant queries can use approved infrastructure signals and return recommendations for engineers.
PlannedSearch your runbooks, incident history, architecture docs. Contextual answers grounded in your own data.
PlannedBrowser-based control panel: server tree, live port status, audit timeline, credential management.
PlannedAI-driven anomaly detection — the platform alerts you before the user does.
PlannedOpsControl is early-stage. We're looking for infrastructure and security teams to pilot controlled access workflows, Vault integration, auditability, and AI-assisted operational analysis.
Based in the Philippines · Self-hosted · On-premise first · Early-stage product