Managed Hermes hosting
The autonomous AI agent platform — deployed, managed, and kept up to date by HostAgentes. Build agents with persistent memory, multi-model routing, and real-time observability without managing infrastructure.
Unlimited agents. Unlimited runs. Persistent memory included. From $9.99/mo.
What is managed Hermes hosting?
Hermes is an open-source AI agent platform built for long-running autonomous agents with native persistent memory, multi-model routing, and production observability. It ships a visual agent builder, a built-in vector store and key-value state layer, team workspaces with role-based access, and conversation-level observability — all in one self-hosted package. Managed Hermes hosting means a third-party provider deploys and maintains a private Hermes instance on your behalf, handling infrastructure so you can focus on building autonomous agents instead of managing Postgres, Redis, vector indices, or state storage yourself.
HostAgentes offers managed Hermes hosting in three tiers: Basic at $9.99/mo (1 vCPU, 2 GB RAM, persistent memory included), Pro at $29.99/mo (2 vCPU, team management, production observability), and Scale at $79.99/mo (4 vCPU, multiple workspaces, target 99.9% uptime SLA). Unlike generic cloud platforms, HostAgentes charges for dedicated compute with no usage caps on agents or runs.
Every HostAgentes Hermes instance comes with pre-configured vector storage and key-value state, PostgreSQL for app state, Redis for caching, and S3-compatible object storage for uploaded documents. SSL is provisioned automatically. Deployment takes under five minutes with no Docker Compose or CLI commands required.
Simple, transparent pricing
No run limits. No per-seat charges. Pay for compute, not usage.
Basic
For experimenting with Hermes and deploying your first autonomous agents.
- 1 vCPU, 2 GB RAM
- 20 GB storage
- Unlimited agents
- Unlimited runs
- Bring your own model keys (BYOK)
- Persistent memory (vector + KV)
- SSL and automatic backups
- Community support
Pro
For teams running production agents with team collaboration and advanced observability.
- Everything in Basic
- 2 vCPU, 4 GB RAM
- 50 GB storage
- Team management + roles
- Advanced agent orchestration
- Production observability dashboard
- Email support
- Automatic zero-downtime updates
Scale
For organizations with multiple high-traffic agent fleets and strict compliance needs.
- Everything in Pro
- 4 vCPU, 8 GB RAM
- 100 GB storage
- Multiple workspaces
- Enterprise SSO
- Priority support
- Target 99.9% uptime SLA
7-day money-back guarantee. Cancel at any time.
The price you see is the price you pay. No renewal surprises, no hidden fees.
The full Hermes stack, managed
Every feature Hermes ships — persistent memory, multi-model routing, visual builder, observability — live on your instance from day one.
Persistent agent memory
Every agent retains context across sessions using a built-in vector store and key-value state. No external Redis or database setup required — memory is provisioned, backed up, and scaled automatically alongside your Hermes instance.
Multi-model routing
Route different agents to different providers — OpenAI, Anthropic Claude, Google Gemini, Mistral, or any OpenAI-compatible endpoint — without vendor lock-in. Switch models per agent or per task with a single configuration change.
Visual agent builder
Design agent workflows in a visual canvas: chain triggers, tools, LLM calls, and memory retrievals without writing code. Export the graph as YAML for version control, or import existing definitions in seconds.
Real-time observability
Monitor latency, token spend, run success rates, and memory retrieval quality per agent from a single dashboard. Set alerts on error rates or cost thresholds and get notified before a small issue becomes an outage.
Team workspaces
Organize agents into isolated workspaces per team or project. Control who can deploy, edit, or view production agents with granular role-based access. Audit logs track every change for compliance reviews.
BYOK by design
Your API keys, your model costs. Hermes connects directly to the model provider from your HostAgentes instance — we never proxy or mark up LLM calls. You see exact token costs in your provider dashboard and in Hermes' observability view.
Popular Hermes agents on HostAgentes
What teams actually build once they have a managed Hermes instance with persistent memory and no run caps.
Customer support agent
Index your help docs and past tickets into persistent memory. The agent answers tickets in real time, escalates to humans when confidence is low, and learns from resolved cases to improve future replies.
Sales automation pipeline
Connect Hermes to your CRM. The agent enriches leads, drafts personalized outreach, schedules meetings, and updates opportunity stages — all while retaining context about previous interactions.
Data processing worker
Ingest CSVs, PDFs, or API feeds on a schedule. The agent validates, transforms, and loads data into your warehouse, reporting errors and retries automatically without human intervention.
Research assistant
Give Hermes a topic and a set of sources. The agent searches the web, reads documents, synthesizes findings into structured summaries, and cites every source with direct links.
IT ops copilot
Monitor logs and alerts via webhook. The agent diagnoses common issues, runs predefined remediation playbooks, and opens tickets in your ITSM tool only when human approval is required.
Content moderation agent
Review user-generated content against your policy documents stored in memory. The agent flags violations, explains the rule triggered, and queues edge cases for human review with full context.
A Hermes agent in 30 lines of YAML
Hermes agents are portable YAML definitions. Export from one instance, import to another, or version-control them alongside your memory configurations. The snippet below is a production-ready support copilot that retrieves from persistent memory, lookups tickets via API, and escalates to Slack when confidence is low.
# Hermes agent configuration — customer support copilot
agent:
name: "Support Copilot"
description: "Answers tickets from help docs and past resolutions"
mode: autonomous
memory:
vector_store: hermes-vectors
state_backend: hermes-kv
retention: 90d
models:
default:
provider: anthropic
model: claude-sonnet-4-5
temperature: 0.3
fallback:
provider: openai
model: gpt-4o
tools:
- name: ticket_lookup
type: api_call
endpoint: "https://api.zendesk.com/v2/tickets"
- name: escalate
type: webhook
url: "https://hooks.slack.com/services/..."
policies:
- rule: "Escalate if confidence < 0.75"
- rule: "Never share billing details"Import this config via Hermes Studio → Import agent. Swap hermes-vectors for your own memory namespace.
From zero to agent in 3 steps
Pick a plan and deploy
Choose Basic, Pro, or Scale. HostAgentes provisions Postgres, Redis, object storage, and the vector + KV memory layer automatically. Your Hermes instance is live with a TLS endpoint in under 5 minutes.
Add your LLM keys and configure memory
Paste your OpenAI, Anthropic, or Gemini API key in Hermes' settings. Then define your agent's memory namespace and upload any documents or policy files. Hermes chunks, embeds, and indexes them automatically.
Build, publish, and monitor
Use the visual builder to design your agent workflow, configure tool calls, and publish your agent as an API endpoint or scheduled worker. Watch latency, run success rates, and memory quality in the observability dashboard.
Why HostAgentes for Hermes?
Hermes' self-hosted stack is complex. HostAgentes tames it so you ship autonomous agents instead of managing containers.
Persistent memory included — no extra bill
Self-hosting Hermes means provisioning and maintaining a separate vector store and state backend. On HostAgentes, persistent memory runs alongside your Hermes instance on the same plan. No extra cloud service, no separate SLA to track.
No per-run pricing trap
Generic agent platforms charge by execution volume: plans cap at a fixed run count and charge overages. HostAgentes charges for dedicated vCPU and RAM. Your agent fleet can run thousands of times a day — the bill stays the same.
All dependencies managed
A production Hermes install requires Postgres, Redis, Celery workers, object storage, a vector DB, and a state backend — all with compatible versions. HostAgentes provisions, updates, and monitors every dependency. You never open a Docker Compose file.
Pair it with Paperclip for orchestration
Hermes excels at autonomous agents with persistent memory. For multi-agent orchestration with governance, budgets, and team collaboration, Paperclip is the right complement. Run both from the same HostAgentes account and let each tool do what it does best.
Frequently asked questions
What makes Hermes different from Dify or Langflow?
How does persistent memory work on Hermes?
Can I use my own LLM keys with Hermes?
Can I run Hermes and Paperclip on the same account?
Is agent memory encrypted at rest?
Deploy your Hermes instance
Persistent memory, multi-model routing, visual agent builder, and production observability — live in under 5 minutes. Unlimited agents and runs from $9.99/mo.