M

Moss Inference

Infrastructure service

by Moss

Private-beta LLM inference that runs on the same host as Moss retrieval so a voice agent's model starts talking sooner; open by access request to teams running long-horizon and real-time agents in production, with no published price.

Find alternatives to Moss Inference

What it does

• Model inference

Sources

Pricing

Not found · 19 Oct 2026

The Moss pricing page and the product page publish no price for Moss Inference.

Description

Official documentation · 19 Oct 2026

"Moss Inference runs optimized LLM inference on the same host as retrieval, so your voice agent starts talking sooner." "Opening to teams running long-horizon and real-time agents in production."

What it does

Official documentation · 19 Oct 2026

"runs optimized LLM inference on the same host as retrieval"

Sold within

Not found · 19 Oct 2026

Read the product page: access is by private-beta request and the page names no parent offering.

Name

Official documentation · 19 Oct 2026

Page heading: "Moss Inference" with a "Beta" label.

URL

Official documentation · 19 Oct 2026

Moss product page titled "Moss Inference · Private Beta", listed under Product in the site menu.

Deployment

Official documentation · 19 Oct 2026

Runs "on the same host as retrieval" in Moss's hosted infrastructure; no self-hosted option is stated.

Interfaces

Official documentation · 19 Oct 2026

The page is a web access-request form; it documents no API, SDK or CLI yet.

Intended for

Official documentation · 19 Oct 2026

"Opening to teams running long-horizon and real-time agents in production."

Also from Moss

• Moss Infrastructure service

  Semantic search runtime for AI agents that indexes documents and answers queries inside the agent's own process, in the browser, on edge runtimes, on devices or in the cloud, with Python and TypeScript SDKs and an MCP server; free developer plan, then $30 and $200 monthly plans plus usage costs.

• Founding Agent AI agent

  Voice AI agent for a company's website that answers visitors from the company's own product, pricing and documentation, labels each visitor's intent and lead score, books demos on the company's calendar and hands high-intent visitors to a live teammate; created and published in the Moss portal and embedded with a React widget.

Open Moss in the directory

Something wrong here? Send a correction — quote this product id: moss-inference.