01 / SUBSTRATE

The substrate for agentic AI.

Memory, orchestration, and tool discovery.
Every agent team rebuilds these badly.
We ship them as infrastructure.

// Beyond the wrapper. Beneath the agent.

02 / THE PROBLEM

Every AI startup is solving the same 5% of the problem.

The prompt layer is done. The 95% — memory that survives sessions, orchestration that respects trust boundaries, tool discovery that doesn't hardcode, multimodal grounding that doesn't hallucinate — is still open. That's what we build.

// Vendor-neutral by design — your stack survives the next model release.

03 / ARCHITECTURE

Not a demo. A platform.

MEMORY

Categorized memory (SGM)

Beyond vector RAG. Memories typed by role, trust, recency.

ORCHESTRATION

Trust-aware orchestrator

Every action carries provenance. Zero-trust by construction.

MCP

MCP-native tool discovery

Agents grow their own capabilities. No hardcoded registries.

VISION

Multimodal grounding

MediaPipe + RT-DETR pipeline. Language tied to perception.

EVENTS

Event-backbone substrate

Async-first. Multi-agent coordination without polling.

SECURITY

Zero-trust primitives

Built into the substrate, not bolted on top.

Vendor-neutral by design — swap any layer

Model
Memory
Voice

// Swap any layer — one env var change

// SUBSTRATE TOPOLOGY

provenance-tracked

Trust-Aware Orchestrator

categorized · typed

SGM Memory

tools grow themselves

MCP Discovery

MediaPipe · RT-DETR

Vision Pipeline

async · multi-agent

Event Backbone

policy at the substrate

Zero-Trust Guardrails

04 / VERTICAL PROOF

SuryaSetu — our next vertical, ready for production.

Solar lead qualification — B2B India market, voice pipeline, real calls. The pipeline is built and validated. We're moving to production deployment now.

Ready for productionVoice pipelineNext vertical — testing

Live pipeline — SuryaSetu

INPUT

Voice call

VAD

Silero

STT

Sarvam

SUBSTRATE

Drevan

LLM

Groq

TTS

Bulbul

TRANSPORT

LiveKit

Barge-in detection · SGM memory across calls · trust-gated qualification decisions

LIVE IN DREVAN

This isn't a roadmap. It's what ships today.

Every feature below is running in production on SuryaSetu — not promised, not demoed, not “coming soon.”

Voice Pipeline

READY FOR PRODUCTION

Real-time voice agent: Silero VAD barge-in detection → Sarvam STT (Hindi/English) → Groq LLM → Bulbul TTS → LiveKit WebRTC. Sub-2s round-trip in field conditions.

Vision Pipeline

INTEGRATED

MediaPipe + RT-DETR wired into agent context. Visual state (objects, faces, scene) flows into the LLM as structured primitives — not raw image bytes.

Inbound & Outbound Calls

READY FOR PRODUCTION

Agents make and receive phone calls via LiveKit WebRTC. SuryaSetu calls leads, qualifies them, logs decisions to SGM — no human in the loop for tier-1 qualification.

SGM Memory — Cross-Call Recall

READY FOR PRODUCTION

Semantic Graph Memory persists intent, outcomes, and consent across every call. A lead who mentioned "roof space" in call 1 is remembered in call 3 — without a database query in the prompt.

LangGraph Orchestration

INTEGRATED

State-machine agent flow via LangGraph. Deterministic qualification branches, escalation nodes, and halt conditions — not a prompt loop hoping the LLM decides correctly.

MCP Runtime Discovery

INTEGRATED

Tools register at runtime via Model Context Protocol. New capabilities appear without redeployment — no hardcoded tool lists, no schema-drift failures.

Screen Automation

RUNNING IN PRODUCTION

Computer-vision-based UI control: EasyOCR + YOLO + template matching. Agents click, type, and verify at pixel accuracy — DPI-aware, with human-like mouse paths and retry logic.

Device Task Engine

RUNNING IN PRODUCTION

Windows system automation: open apps, move/copy/delete files, search directories, read system state. Path-restricted, extension-whitelisted, with an async approval queue for destructive operations.

WhatsApp & Email

INTEGRATED

Agents send WhatsApp messages and emails via fuzzy contact resolution — no exact name required. Confidence-scored matching before dispatch, guardrail-gated by default.

Google Calendar

INTEGRATED

Create, list, check availability, daily briefing, delete events via OAuth2. Natural-language scheduling wired into agent context — "find me a free slot Thursday afternoon" works.

Proactive Agent

INTEGRATED

Unsolicited execution without a user prompt — scheduled briefings, weather checks, delegated task monitoring. Cron-like scheduler with a daemon runtime and event-log replay.

05 / ENGINEERING LOG

Open notebook. Closed source.

Every major architectural choice — documented, dated, defended.

Voice pipeline on Drevan: Silero VAD → Sarvam STT → Groq → Bulbul TTS

End-to-end voice agent on the substrate. Latency profile, barge-in detection, LiveKit WebRTC — and why we chose Sarvam over Deepgram for Hindi.

LangGraph + MCP phase 1: ComponentFactory wired to runtime tool discovery

Dynamic tool registration at runtime — no hardcoded imports. The architecture decision that will matter when you add your 12th tool.

SGM schema, category taxonomy, and the consent model — design notes

How we designed the Semantic Graph Memory layer: write guards, cross-session retrieval, and why flat vector DBs fail at agent memory.

// Next post: 2026.06.01

05 / EARLY ACCESS

I talk to three builders a month.

Early access is not a waitlist.

I onboard three builders a month. You bring a real agent product, a real production problem, and 30 minutes a week. I bring the substrate and ship to your specific constraints.

No demo calls. No sales deck. If you're building with agents and hitting the substrate problems — memory that breaks, orchestration that drifts, tool discovery that's hardcoded — write directly.

AB

Abhitesh Bhardwaj

FOUNDER · DREVAN

Building agentic systems in production at a venture-backed AI company. Shipping Drevan + SuryaSetu on the side.

Write to

abhitesh@drevan.io

Include: what you're building · where orchestration/memory is breaking · stack

We don't do RAG-as-a-service. We don't do chatbots.