Runtimes

Runtime support#

ClawMetry reads 30 agent runtimes. 3 are free in the open-source package; 27 unlock on a paid tier. Every one of them is read from its own on-disk (or, for two, remote) store — there is no SDK to import and nothing to instrument.

Aider paid

AI pair programmer

Antigravity paid

Agentic IDE and CLI

Claude Code paid

Terminal coding agent

Cline paid

Agentic coding CLI and extension

Codex paid

Terminal coding agent

Cursor paid

AI IDE (editor + CLI)

Deep Agents paid

LangGraph agent harness

DeepSeek Harness paid

Agent harness

Devin paid

Autonomous engineer (CLI)

Exo paid

Self-improving agent harness

Gemini CLI paid

Terminal coding agent

GitHub Copilot paid

Terminal coding agent

Goose free

Open-source agent (Rust CLI + desktop)

Grok paid

Terminal coding agent

Grok Bot paid

Cloud agent, desktop client

Hermes paid

OpenClaw-descendant agent

Kimi paid

Terminal coding agent

Lovable paid

Cloud app builder

n8n paid

Workflow automation with AI agent nodes

NanoClaw paid

Containerised agent runtime

NemoClaw free

Governed agent runtime

OpenClaw free

Personal agent gateway

opencode paid

Terminal coding agent

OpenHands paid

Autonomous software agent

OpenWorker paid

Generalist desktop worker

Pi paid

TypeScript coding agent

PicoClaw paid

Minimal Go agent

QM paid

Multiplayer agent harness

Qwen Code paid

Terminal coding agent

Replit Agent paid

Cloud app builder

The full list#

Runtime	id	Category	Vendor	Tier
Aider	`aider`	AI pair programmer	Aider AI	Paid
Antigravity	`antigravity`	Agentic IDE and CLI	Google	Paid
Claude Code	`claude_code`	Terminal coding agent	Anthropic	Paid
Cline	`cline`	Agentic coding CLI and extension	Cline	Paid
Codex	`codex`	Terminal coding agent	OpenAI	Paid
Cursor	`cursor`	AI IDE (editor + CLI)	Anysphere	Paid
Deep Agents	`deepagents`	LangGraph agent harness	LangChain	Paid
DeepSeek Harness	`deepseek_harness`	Agent harness	DeepSeek	Paid
Devin	`devin`	Autonomous engineer (CLI)	Cognition	Paid
Exo	`exo`	Self-improving agent harness	Exo Harness	Paid
Gemini CLI	`gemini_cli`	Terminal coding agent	Google	Paid
GitHub Copilot	`copilot`	Terminal coding agent	GitHub	Paid
Goose	`goose`	Open-source agent (Rust CLI + desktop)	Block	Free
Grok	`grok`	Terminal coding agent	xAI	Paid
Grok Bot	`grok_bot`	Cloud agent, desktop client	xAI	Paid
Hermes	`hermes`	OpenClaw-descendant agent	Nous Research	Paid
Kimi	`kimi`	Terminal coding agent	Moonshot AI	Paid
Lovable	`lovable`	Cloud app builder	Lovable	Paid
n8n	`n8n`	Workflow automation with AI agent nodes	n8n	Paid
NanoClaw	`nanoclaw`	Containerised agent runtime	Nanoco	Paid
NemoClaw	`nemoclaw`	Governed agent runtime	NVIDIA	Free
OpenClaw	`openclaw`	Personal agent gateway	OpenClaw	Free
opencode	`opencode`	Terminal coding agent	SST	Paid
OpenHands	`openhands`	Autonomous software agent	OpenHands	Paid
OpenWorker	`openworker`	Generalist desktop worker	Andrew Ng	Paid
Pi	`pi`	TypeScript coding agent	Earendil	Paid
PicoClaw	`picoclaw`	Minimal Go agent	Sipeed	Paid
QM	`qm`	Multiplayer agent harness	Y Combinator	Paid
Qwen Code	`qwen_code`	Terminal coding agent	Alibaba	Paid
Replit Agent	`replit`	Cloud app builder	Replit	Paid

Two runtime ids exist in the entitlement list but share another runtime's read path: grok_bot is read through the Grok adapter. Adapters also ship for Lovable and Replit.

What "supported" means here#

A runtime being on this list means an adapter exists that reads its real store and maps it into ClawMetry's unified event schema. It does not mean every runtime exposes the same things — they genuinely do not, and the difference matters:

Cline records dollars, a per-message token split, process ids, exit codes,

sub-agent lineage, real cron schedules with per-run cost, and even the tool calls it rejected. It is the richest store in the matrix.

PicoClaw's message struct has no token field at all. ClawMetry reports

zero tokens and does not advertise cost for it, because the data is not there.

Cursor's backend is proprietary, so token counts come only from a hosted

VM's loopback usage log — which is why a Cursor row can show real cost on a managed node and none on a laptop.

Kimi persists a complete token breakdown but never the model id, so cost

is marked estimated with the model resolved from config.

What each runtime exposes has the comparison table. Each runtime's own page has the exact paths, environment overrides and traps.

Adding a runtime that is not here#

Three options, in increasing order of effort:

OpenTelemetry. If the runtime can export OTLP, point it at ClawMetry's

receiver. This is the supported path for n8n on Postgres, hosted CI agents and anything without a local store. → OpenTelemetry ingest

The ingest API. Push runs and events over HTTP from code you control.

→ Custom runtime ingest

The interceptor. Two lines in a Python process capture its LLM calls,

cost, tokens, latency and errors, attributed to a named source. → Track any SDK agent

If you want a first-class adapter for a runtime that writes to disk, open an issue with the store layout — that is the part that takes the work.

Runtime ids#

The id column is what appears everywhere machine-readable: the entitlement API, clawmetry runtimes --json, the runtime filter on query shapes, session id prefixes, and per-runtime environment overrides like CLAWMETRY_NOPROG_TOOLS__CODEX. Common alternate spellings are accepted as aliases — claude-code and claudecode both resolve to claude_code.

Edit this page View as Markdown Report an issue