Hermes LoopHermes Loop
Runtime

Hermes control room

Live status of the connected Hermes runtime, configured capabilities, recent requests, and direct entry points to the marketplace, memory inspector, tool sandbox, and command inbox.

Hermes connected
checked 1:58:21 AM · 102ms · model nousresearch/hermes-4-70b
API base
https://openrouter.ai/api/v1
Configured
yes
Last error

Control surfaces

Capabilities

Chat completions
chat
available
hermes
POST /chat/completions verified
Tools
tools
available
agent foundry
Sandboxed execution layer in Hermes Loop — see /hermes/tools
Memory
memory
available
agent foundry
Backed by Hermes Loop MemoryItem — see /hermes/memory
Skills marketplace
skills
available
agent foundry
Local + marketplace skills in /hermes/skills
Cron / schedules
schedules
available
agent foundry
Persistent ScheduledMission + RUN_SCHEDULE jobs
File access
files
available
agent foundry
Governed terminal_exec can inspect workspace files through allowed read-only commands.
Browser automation
browser
available
agent foundry
browser_qa_audit uses real Playwright/Chromium with screenshots and console capture.
Search / vision / media
media
available
agent foundry
web_search, vision_analyze, image_generate, and text_to_speech are wired and provider-gated.
Terminal / Python runtime
runtime
available
agent foundry
terminal_exec and python_rpc run server-side with policy gates, timeouts, and approvals.

Test chat

Sends a one-shot probe via the Hermes adapter. Falls back to demo mode if Hermes is unreachable. Logged in the request log above.

Recent Hermes requests (0)

No requests yet. Run a mission, send a test chat, or trigger a job to populate the log.