Epsilla

Company Overview

As of July 20, 2026: Epsilla is a RAG-based AI agent platform targeting enterprise buyers, with plans running from free to $2,000/month before usage costs stack on top.

The platform covers 16 stated capabilities including RAG and multi-tenancy, with SSO, SAML, and RBAC available at higher tiers. Vector storage at $18/GB/month and per-message charges at $3 per 100 messages mean costs can climb quickly for data-heavy deployments.

The record, by dimensionevery dimension is its own page →

Pricing

Enterprise, sales-led

Five tiers from $0 to $2,000/month, plus a contact-sales Enterprise tier, with six usage-based add-ons layered on top. A buyer on the Professional tier at $249/month who adds seats ($12/seat/month), knowledge bases ($2/KB/month), and meaningful vector storage ($18/GB/month) will see the real bill diverge from the headline price quickly. White-label removal costs an additional $39/month.

Full pricing read →

Features

RAG platform

Epsilla states 16 capabilities anchored by RAG and multi-tenancy. Multi-tenancy matters for enterprise buyers building customer-facing agents, but no independent feature audit is available to verify depth or completeness of any individual capability.

Full read →

Security

Baseline security

SSO, SAML, and RBAC are the three stated controls. That covers basic enterprise procurement requirements, but there are no stated certifications (SOC 2, ISO 27001, or similar) in the provided signals, which will be a gap for buyers in regulated industries.

Full read →

Reliability

Dependable

A 99.9% uptime SLA is stated on the site, which allows for roughly 8.7 hours of downtime per year. No public uptime history or status page data was provided to corroborate that figure.

Full read →

Hiring

4 open roles

Four open roles span engineering, design, and marketing. That is a small active headcount footprint; buyers evaluating long-term vendor stability should note the company appears to be in an early growth phase. Remote work policy is not stated.

Full read →

Stack

Detected + stated stack

The live site runs Next.js, Tailwind CSS, and Vercel, with React confirmed in both observed and stated sources. Job postings additionally call out Python, PyTorch, and TensorFlow, indicating a machine-learning-oriented backend that site fingerprinting cannot see. The Vercel deployment is worth noting for buyers with data-residency requirements, as infrastructure control may be limited.

Full read →

Positioning

AI Agent Platform

The homepage headline is "Build Vertical AI Agents Without Engineering Overhead," aimed squarely at enterprise buyers who want to deploy AI agents without internal infrastructure work. The Agent-as-a-Service framing puts the burden of uptime and scaling on Epsilla, which makes the unverified SLA more consequential.

Full read →

Location

Jersey City, NJ

Epsilla, Inc. is headquartered in Jersey City, NJ. Work model is not stated, which is relevant context given the four open roles and the small team size.

Full read →

Legal

Epsilla, Inc.

The registered entity is Epsilla, Inc. Incorporation state, governing law, arbitration terms, class-action waiver, and auto-renewal policy are all unstated in available signals. Enterprise buyers should expect to negotiate these terms directly before signing.

Full read →

Change History30 dated events on the record

Jul 2026

Positioning changed: Target Segment

Dec 2025

Pricing changed: Team Member Seat usage rate · high significance

Dec 2025

Pricing changed: Remove Powered By Epsilla usage rate · high significance

Dec 2025

Feature added: White Labeling

Dec 2025

Pricing changed: Remove Powered By Branding usage rate · high significance

Dec 2025

Pricing changed: AI Concierge tier monthly price · high significance

Dec 2025

Pricing tier added: AI Concierge tier name · high significance

Get this record via API / MCPevery field dated and sourced

RESTopen tier

GET https://api.bixel.com/v1/companies/epsilla.com

{
  "data": {
    "company": {
      "name": "Epsilla",
      "domain": "epsilla.com",
      "categories": [
        "vector-databases"
      ]
    },
    "dimensions": {
      "reliability": {
        "facts": [
          {
            "key": "reliability.sla_pct",
            "value": 99.9,
            "provenance": "company_stated",
            "as_of": "2026-07-20",
            "source_url": "www.epsilla.com/sla"
          },
          {
            "key": "hiring.open_roles",
            "value": 4,
            "provenance": "company_stated",
            "as_of": "2026-07-20",
            "source_url": "www.epsilla.com/careers"
          },
          {
            "key": "stack.tech.nextjs",
            "value": true,
            "provenance": "observed",
            "as_of": "2026-07-20",
            "source_url": "www.epsilla.com"
          }
        ]
      },
      "…": "…"
    }
  }
}

MCPfor agents

# any MCP client (Claude, agents)
const record = await bixel.get_company({ domain: "epsilla.com" })

# returns the whole record above,
# each value with its source_url + as_of,
# ready to reason over

Build on the company record. One key, REST + MCP, every signal dated and sourced back to the page it came from.

How Bixel reads this

Every read above is derived from public signals, each sourced and dated, and kept honest about provenance: what the company states on its own pages and job posts (pricing, security, careers, positioning, its backend stack) versus what Bixel independently detects (technologies, infrastructure). Where a stated claim is also detected we mark it verified; where we only have the claim, we say so. Bixel infers posture: how it monetizes, how mature it is, where it's heading. It does notclaim private financials it can't observe. Where signals are thin, the confidence says so.

Public record, read from companies' own pages and boards. Every fact dated and sourced; provenance (observed vs company stated) shown inline.