comparison · named names

Named products. Checkable claims.

Self-hosted agents people actually evaluate: Smartopol, Hermes, OpenClaw, NanoClaw, IronClaw, OpenJarvis, plus coding agents such as Claude Code and Grok Build. Where Smartopol leads, where others lead, and what is a different category entirely.

free to use · closed source by design · benchmarks you can run with smartopol bench

§ 01 · one-screen summary

Every row is meant to be verifiable.

capabilitysmartopolhermesopenclawnanoclawironclawopenjarvis
Self-hostedyesyesyesyesyesyes
Nightly reflectioncortex dreamingskill rewrites
Hierarchical memoryMemory Pyramidkeyword / notesbasicbasicencrypted KVbasic
Cross-channel identitypairing codespartialper-channel
Channels (shipped)Discord · TG · Web · WS · Terminalmany claimedmany claimedfewlimitedlocal
Model-agnosticGrok · Claude · GPT · Qwen · Kimi · OllamaClaude-firstClaude-firstClaude-firstClaude-firstClaude-first
Sandboxed pluginssigned, isolatedscriptsskills hubbasicbasicbasic
Memory write scanningthreat patterns + fencepartialpublic CVEs 2026yes
OS sandboxkernel-enforcedclaimedweak defaultscontainerzero-trust
Audit logchainedclaimed
Self-updatesigned binarypip -Umanualmanualmanual
Multi-user RBACyesper-agentyes
Published benchmarkssmartopol bench
Licensefree to use · closed sourceMITMITMITcommercialMIT
Pricingfree to usefreefreefreepaidfree
Commercial supportyesyes
Smartopol lead present- absentwarn caveat
§ 02 · vs Hermes

Hermes Agent (Nous Research)

Closest positioning rival: autonomous agent on your server that remembers and grows over time. Same category as Smartopol.

Where Hermes leads

  • Channel breadth: many messaging adapters claimed
  • Deployment shapes: Docker, VPS, remote backends
  • Public community skills: open hub culture

Where Smartopol leads

  • Memory Pyramid: hierarchical hybrid recall, not only keyword notes
  • Nightly reflection: bounded, auditable consolidation
  • Hardening defaults: write scanning, fenced recall, sandbox plugins
  • Signed self-update: binary swap, not only package managers
  • Published bench path: smartopol bench
honest read

Both free to use in spirit; Smartopol is closed source, Hermes is MIT.

Pick Hermes for open community + channel sprawl. Pick Smartopol for memory depth + operator hardening.

§ 03 · vs OpenClaw

OpenClaw

Strong on channels and skill marketplace. Weaker on default security posture.

Where OpenClaw leads

  • Channel count
  • Skill marketplace size
  • Fast setup narrative

Where Smartopol leads

  • Safer defaults: smaller curated surface vs large unaudited skill hubs
  • Memory system: pyramid, not only persistent context bags
  • Multi-provider design: not Claude-only
  • Commercial path: free to use binary + support when needed
honest read

If you want maximum marketplace size and accept marketplace risk, OpenClaw feels richer. If you want a smaller, sandboxed, operator-grade default, Smartopol is the point.

§ 04 · vs NanoClaw

NanoClaw

Minimal, auditable, security-focused alternative to larger OpenClaw-class stacks.

Where NanoClaw leads

  • Tiny surface: easy to read end to end
  • Container isolation
  • Cheap hardware bill

Where Smartopol leads

  • Memory Pyramid depth
  • Skills + plugin system
  • Cross-channel identity
  • Multi-user service model
honest read

NanoClaw optimizes for smallest auditable surface. Smartopol optimizes for capable private agent with responsible defaults. Different products.

§ 05 · vs IronClaw

IronClaw

Enterprise zero-trust framing. Paid product path. Compliance vocabulary.

Where IronClaw leads

  • Enterprise sales motion
  • Compliance checklist language
  • Encrypted memory defaults

Where Smartopol leads

  • Free to use binary: IronClaw is paid-only
  • Faster engineer setup: install path, not RFP first
  • Model neutrality
  • Support available without locking the product behind a paywall to start
§ 06 · vs OpenJarvis

OpenJarvis

Local-first personal agent framing. Earlier-stage than the rest of this list. Smartopol is also local-first, and adds Memory Pyramid, channels, plugins, multi-user, and a published bench path.

§ 07 · vs Claude Code

Claude Code is a hand. Smartopol is a long-lived brain.

Where Claude Code leads

  • First-party Anthropic product
  • IDE integrations
  • Continuous model improvements from Anthropic

Where Smartopol leads

  • Channels: Discord, Telegram, web, WS, terminal with shared identity
  • Provider neutrality
  • Memory Pyramid
  • Always-on multi-user service
  • Can drive Claude Code as a tool: complementary pattern
honest read

Not substitutes. Typical setup: Smartopol as the long-lived agent; Claude Code (or Grok Build) as a focused coding hand.

§ 08 · vs Grok Build · Qwen Code · Kimi Code · Aider

Coding agents are a different category.

Grok Build is xAI's coding agent product (Grok is the model family).Qwen Code and Kimi Code are coding agent CLIs. Aider and similar tools are repo-first editors.

Smartopol Agent is a general-purpose companion with memory, channels, scheduling, and multi-user service. It can code, and it can also keep life and ops context across weeks. Memory Pyramid is especially useful for long Grok Build sessions that need context across restarts.

If you only want a coding agent, use Grok Build, Qwen Code, Kimi Code, Claude Code, or Aider. If you want one long-lived agent that can use those as tools, use Smartopol.

§ 09 · hard questions

Questions a serious buyer asks.

Will my data leave my machine?

Only if you send prompts to a cloud model provider you configured. Memory defaults to local storage. Telemetry is off by default.

Is it free forever?

No forever claim. The binary is free to use under the license on this site. Closed source by design. Terms can evolve with a public letter if they change.

Where is the source?

Not a public forge today. See /why-closed-source. Source review under NDA is case by case.

Do you sell my data?

No. No ads. No data brokering. No mandatory telemetry.

§ 10 · notes

Traceability.

Private beta

Invitations go out in waves. Free to use. Closed source by design.

No nurture sequence.