Hermes Business
by Nous Research
Hosted Hermes Agent for whole teams: one shared balance, per-member spend caps and a shared skills library
Hermes Business is Nous Research's team edition of the open-source Hermes Agent. Every member gets a hosted agent reachable from Slack, Discord, Telegram, WhatsApp, the terminal or desktop apps, billed from one central balance with per-member caps. It is for teams that want a shared, self-improving agent without running their own servers.
Hermes Business is the commercial team tier of Hermes Agent, the MIT-licensed, Python agent that Nous Research open-sourced in July 2025. The GitHub repository had about 252,000 stars and 54,000 forks on October 9, 2026, and Nous says the agent has been cloned more than 24 million times and drives roughly 2.5% of global AI token usage. Hermes Agent keeps agent-curated memory across sessions, writes its own skills from experience in the open agentskills.io format, runs scheduled jobs, spawns subagents, supports MCP servers, and can execute in seven terminal backends including Docker, SSH, Daytona, Modal and Vercel Sandbox. It reaches users through more than 20 messaging platforms from one gateway. Hermes Business, announced on X in mid-September 2026, moves that agent onto Nous infrastructure for a team: owners invite colleagues by email, each member gets a hosted agent, everyone draws on one central balance with per-member spend caps, and any skill a member creates is published to a team library that every agent can use. Members can call any model on the Nous inference API or bring their own provider keys, and each team runs in an isolated tenant. A separate Hermes Enterprise tier, open only for registration of interest, will run on customer infrastructure with SSO, SLAs and assisted onboarding. On October 7, 2026 Nous confirmed a $90 million Series B at a $1.5 billion valuation led by Robot Ventures to push into enterprise, with about $36 million in annualised revenue reported in mid-September.
An engineering, operations or AI-enablement lead at a small or mid-sized technical team that already experiments with Hermes Agent and wants it hosted, shared and budget-capped.
Every team member gets a persistent agent in the chat tools they already use, with skills one person builds becoming available to everyone and spend capped per person.
At a Glance
- Category
- AI Agents & Orchestration
- Pricing
- Usage-based, Freemium, Contact for pricing
- Target Market
- Enterprise Developers, Engineering Managers, Operations Teams, AI Platform Teams
- Deployment
- Cloud-first, Open-source
- Founded
- 2023
Key Features
- ✓Hosted agent per member
Each invited colleague gets their own hosted Hermes agent on Nous infrastructure, with no VPS or container for the team to run.
- ✓Central balance with spend caps
One team balance funds all usage, owners set per-member caps, and the portal shows every dollar spent by whom.
- ✓Shared skills library
Skills a member's agent creates are published to the team library and become available to every agent, so automations compound.
- ✓Team Gateway channels
Agents are reachable from Slack, Discord, Telegram and WhatsApp plus terminal, desktop apps and cloud, meeting people where they work.
- ✓Model choice
Members can use any model on the Nous inference API or plug in their own provider keys, avoiding single-vendor lock-in.
- ✓Persistent memory and self-written skills
Inherited from open-source Hermes Agent: cross-session memory, FTS5 recall and skills the agent writes and refines from experience.
- ✓Isolated tenant
Each Business team runs in its own isolated tenant on Nous infrastructure, with an Enterprise path to on-prem or private cloud.
Capabilities
Use Cases
- •Team operations assistant
An ops team gives every member an agent in Slack that drafts reports, runs lookups and reuses skills colleagues already built.
- •Scheduled status reporting
Agents run cron jobs overnight and deliver a morning summary to a chat channel, replacing manual daily status collection.
- •Graduating from self-hosted Hermes
A team that ran Hermes Agent on its own VPS moves to hosted agents and one budget, dropping server upkeep.
- •Controlled model experimentation
A team tries several models through one harness while per-member caps keep total experimentation spend inside a fixed budget.
Ideal For
Best For
- ✓Technical teams already running self-hosted Hermes Agent who want to stop managing servers
- ✓Sharing reusable agent skills across a team so automations built once spread to everyone
- ✓Agents that work from Slack, Discord, Telegram or WhatsApp rather than a separate web app
- ✓Model-agnostic teams that want one harness across Nous-hosted models and their own provider keys
- ✓Scheduled, recurring agent jobs such as daily reports delivered to a chat channel
Not Ideal For
- ✗Regulated workflows (legal, healthcare, finance, HR) that need audit logs, approvals and certifications; the Business page lists no SOC 2 or other compliance attestations
- ✗Organisations that need SSO, SLAs or data kept on their own infrastructure today, since those sit in the Enterprise tier that is still registration-only
- ✗Predictable, fixed-sequence processes that a workflow tool such as Zapier, Make or n8n handles more cheaply and deterministically
- ✗Non-technical teams without an owner to scope permissions and review early runs; independent guides call the underlying agent experimental
Integrations
Deployment
Market Analysis
Pros
- ✓Very large open-source community behind the hosted product, with the repository updated as recently as October 9, 2026
- ✓Per-member spend caps and one balance make agent costs easy to bound
- ✓Agents live in existing chat tools across more than 20 platforms
- ✓MIT licence leaves a self-hosting exit if the hosted tier does not fit
Cons
- ✗No published compliance certifications for the Business tier, and SSO and SLAs are Enterprise-only and not yet shipping
- ✗Hacker News users call the CLI slow to start and over-engineered, and one widely shared post advised against trusting it with confidential data
- ✗The project has faced plagiarism allegations over its self-evolution design and criticism for pre-installing a Polymarket skill
- ✗Repository maintenance load is heavy: about 47,900 open issues and pull requests on October 9, 2026
Pricing
Nous Portal Free (individual)
$0
- ✓Pay-as-you-go access to 300+ models
- ✓No monthly credits or Tool Gateway
Nous Portal Plus (individual)
From $20/mo
- ✓$22 monthly credits
- ✓Hosted tool access
- ✓Higher rate limits
Hermes Business
Usage-based
- ✓Shared team balance
- ✓Per-member spend caps
- ✓Hosted agent per member
- ✓Shared skills library
Hermes Enterprise
Contact for pricing
- ✓Runs on your infrastructure
- ✓SSO
- ✓SLAs
- ✓White-glove onboarding
Hermes Business publishes no per-seat price: the team pre-funds one central balance that model and tool usage draw down, with owner-set per-member caps. Individual Nous Portal plans run $0, $20, $100 and $200 a month with slightly larger credit allowances (per Hostinger's guide, updated September 15, 2026). The open-source agent itself is free under MIT; Enterprise is quote-only.
Security & Compliance
Connect
Sources
This page was written from 8 sources, 7 on domains other than portal.nousresearch.com.
- 1.portal.nousresearch.com — businessvendor
- 2.hermes-agent.nousresearch.com — docs
- 3.techcrunch.com — nous research confirms it hit 1 5b valuation launches ai age
- 4.api.github.com — hermes agent
- 5.hn.algolia.com — 48419000
- 6.hn.algolia.com — search
- 7.aslanintelligence.com — hermes agent for business
- 8.hostinger.com — hermes agent cost
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Lindy
No-code AI agents that handle email, meetings, CRM updates and phone calls across 1,000+ business apps
Dust
Multiplayer AI agent platform where teams build, share and govern company-wide AI agents on their own data
SuperNinja Enterprise
24/7 AI employees on open-weight models, deployed in your own cloud for a fixed annual fee
Capacity
AI-native customer experience platform: omnichannel AI agents, real-time agent assist and automated QA on one knowledge layer