Moss

Real-time semantic search for Conversational AI

usemoss.dev

HQ: San Mateo, CA, USA

1-10

B2B

Company research

Public company, workplace, funding, and market signals

Updated Jul 28, 2026

Overview

Moss (legal entity: InferEdge Inc.) is a YC Fall 2025 developer-infrastructure startup building a real-time semantic-search runtime for conversational AI, voice agents, copilots, and multimodal applications. Its approach keeps retrieval local/on-device or close to the application, avoiding vector-database network round trips.

Primary product

Moss real-time semantic-search/retrieval runtime and SDKs for browser, mobile, edge, server, and cloud applications; it offers sub-10 ms lookups, hybrid semantic-plus-keyword search, metadata filtering, instant index updates, built-in or customer-supplied embeddings, and optional cloud synchronization.

Founded

2024

Headquarters

San Francisco Bay Area, California, United States

Team size

1–10 (LinkedIn snapshot); other YC job pages report 5 team members, so public counts vary by date

Work style

Flexible

Industry

Technology, Information and Internet

Sub-industry

AI infrastructure; semantic search and retrieval; developer tools

Offices

San Francisco Bay Area, California, United States

San Francisco, California, United States

Business model

Hosted B2B SaaS/API with subscription tiers

Usage-based cloud retrieval and synchronization

Open-source core/SDK-led developer adoption

Enterprise contracts with custom scale, onboarding, SLA, SSO, and compliance support

Funding

Stage

Seed / YC-backed

Total raised

$500K

Latest round

Seed · Sep 2025

Latest amount

$500K

Seed

Sep 2025 · Y Combinator

$500K

People & investors

Leadership

Sri Raghu Malireddi

Founder & CEO

Harsha Nalluru

Co-founder & CTO

Sandra Chang

Founding Designer

Ashvath Suresh Kumar

Founding Product Engineer

Investors

Work & culture

Public hiring and founder materials emphasize technical rigor, ownership, performance engineering, low-latency systems, open-source development, and developer experience. The team describes itself as focused on eliminating retrieval latency in conversational AI; specific benefits are not publicly documented.

Work style

Flexible

Visa sponsorship

Limited

Compensation

Public postings show $60,000–$300,000 salary plus 0.10%–0.50% equity for Software Engineer roles, and $120,000–$150,000 plus 0.05%–0.15% equity for Founding Customer Success. Sponsorship is explicitly offered for at least the SDK and backend engineering roles; the customer-success posting is restricted to US citizen/visa-eligible candidates.

Product & market

Pricing

Developer: free with $5/month credits and unlimited local queries; Hobbyist: $30/month; Start-up: $200/month; Enterprise: custom pricing. Hosted features, sync, session replays, concurrency, support, SLA, SSO, SOC 2/HIPAA support vary by tier.

Differentiators

—Sub-10 ms retrieval and reported 5.4 ms P99 benchmark
—Local-first, in-process retrieval that avoids network round trips
—Runs across browser, mobile, edge, server, and cloud from a common runtime/API
—Rust and WebAssembly implementation
—Hybrid semantic plus keyword retrieval with instant updates and metadata filtering
—Privacy-by-architecture/on-device operation with optional cloud sync
—Open-source SDK/runtime and broad integrations
—Founding Agent extends the retrieval infrastructure into an early-access voice website agent

Technology

Rust

WebAssembly

JavaScript/TypeScript

Python

Swift

Kotlin/Android

C/C++ and FFI

Hybrid semantic search and BM25 keyword search

On-device/in-browser indexes

Built-in transformer embeddings

BYOE (Bring Your Own Embeddings)

AWS

Azure

Google Cloud

Metadata filtering

LangChain

LlamaIndex

DSPy

Pipecat

LiveKit

Customers

Microsoft

EPAM

UC Berkeley

Carnegie Mellon

Podium

Stanford

Pipecat/Daily.co and LiveKit are named integration or design-partner relationships, not necessarily paying customers

Six enterprise design partners and three paying customers were reported by YC; individual names were not disclosed

Competitors

Pinecone

Qdrant

ChromaDB

Other hosted vector databases

Traditional RAG retrieval stacks

Scale & contact

Estimated monthly visits

5.3K

Traffic estimate as of Aug 2026

contact@moss.dev

Signals & risks

Other

Jun 2026

Founding Agent, a new voice-agent product for company websites, was announced in early access/waitlist rather than as a generally available product; adoption and product-market fit for this adjacent offering remain unproven.

Source

Other

Apr 2026

The repository underwent a broad rebrand from InferEdge to Moss, including package, documentation, and domain references. This appears to be a naming transition, not a shutdown or acquisition.

Source

Latest news

5,000,000 production voice minutes on Moss (YC F25)

Moss (YC F25) LinkedIn · Jul 2026

Company milestone post reporting Moss reached 5 million production voice minutes.

Update: Moss (YC F25) just crossed 1 million minutes of usage

Moss (YC F25) LinkedIn · Jun 2026

Company milestone post reporting Moss crossed 1 million minutes of usage.

We Built a Voice AI Agent for Our Website. Then Other Companies Started Asking for It.

Moss · Jun 2026

Official product launch post for Founding Agent, a voice AI agent for company websites that answers visitor questions and qualifies prospects.

Excited to share that we are now live on Launch YC!

Moss (YC F25) LinkedIn · Nov 2025

Company announcement that Moss went live on Launch YC and reiterated the product’s real-time retrieval positioning.

Launch YC: Moss: Real-time Semantic Search for Conversational AI

Y Combinator · Oct 2025

Official launch post describing Moss’s sub-10ms semantic search runtime, its use for conversational AI, and early traction with enterprise design partners and paying customers.

Jobs and careers at Moss

31 jobs at Moss

Pricing & Deal Desk Manager (m/f/d)

Remote

Commercial

Berlin; Berlin, Berlin, Germany

1w ago

Remote

Full-time