Maturity taxonomy

the model behind the scorecard

Surfaces > categories > capabilities > evidence.

50 surfaces grouped into 4 families, with every category tied back to canonical docs and QA coverage IDs.

Browse product areas / Open detailed taxonomy / View scores

How to read this page

A surface is a product area such as Gateway runtime, Discord, or the macOS app. Each surface contains categories, and each category contains the capability-level checks that QA scenarios cover. Use the scorecard for release-level judgment; use this page to inspect the model underneath it.

Maturity levels

M0PlannedDirection is known, but no supported user path exists.Promotion: Design issue, owner, and target surface exist.

M1ExperimentalImplemented behind caveats, flags, source builds, or maintainer-only flows.Promotion: Maintainer can run the scenario from current main.

M2AlphaReal users can try it, but breaking changes and incomplete UX are expected.Promotion: Documented setup, basic tests, known caveats, and at least one real-environment proof.

M3BetaPublic path exists and the main workflow is usable with bounded caveats.Promotion: Install/update docs, regression tests, support runbook, and successful scenario proof across the expected environment.

M4StableRecommended path for normal users. Failures are treated as regressions.Promotion: Release gate, doctor/troubleshooting path, broad docs, and repeated real-world proof.

M5ClawesomePolished, delightful, well-instrumented, and competitive with the best comparable workflow.Promotion: Stable plus user scorecard pass across representative users.

Product areas

Core
Platform
Channel
Provider and tool

Details

Core

CLI - M4 Stable - 7 areas

Normal setup and repair paths are documented across install, CLI, and gateway docs. Platform-specific Windows paths are tracked in the Windows via WSL2 and Native Windows rows.

Coverage Experimental - 2%Quality Stable - 83%Completeness Stable - 90%Partial - 6

Gateway runtime - M4 Stable - 13 areas

Core architecture, auth, pairing, protocol docs, daemon docs, and CLI runbooks are broad and current.

Coverage Experimental - 3%Quality Stable - 81%Completeness Stable - 89%Partial - 12

Agent Runtime - M3 Beta - 9 areas

Main loop, models, provider routing, and tool streaming are first-class, but provider behavior shifts weekly and needs scenario proof per release.

Coverage Experimental - 2%Quality Beta - 78%Completeness Beta - 79%Partial - 6

Session, memory, and context engine - M3 Beta - 9 areas

Strong docs and active implementation. Maturity depends on transcript durability, compaction quality, and cross-client parity.

Coverage Experimental - 0%Quality Beta - 77%Completeness Beta - 79%Partial - 6

Channel framework - M3 Beta - 8 areas

Many channels share Gateway delivery and routing contracts, but channel behavior varies by upstream API and account-policy constraints.

Coverage Experimental - 0%Quality Beta - 76%Completeness Beta - 79%Partial - 5

Observability - M3 Beta - 5 areas

OTel, Prometheus, logging, and diagnostics docs exist. Needs a public “what operators should look at first” maturity pass.

Coverage Experimental - 6%Quality Beta - 75%Completeness Beta - 79%Partial - 3

Gateway Web App - M3 Beta - 6 areas

Web UI is documented with pairing, chat, PWA, Talk, push, and remote Gateway flows. Promote after cross-browser and mobile-PWA scorecards.

Coverage Experimental - 0%Quality Beta - 74%Completeness Beta - 79%None

Plugins - M3 Beta - 9 areas

Broad docs and strong internal runtime evidence exist across manifests, discovery, loading, provider/tool architecture, and approval boundaries. Keep the row at beta until public SDK API/subpaths and external distribution proof are stronger.

Coverage Experimental - 2%Quality Beta - 72%Completeness Beta - 79%Partial - 7

Security, auth, pairing, and secrets - M3 Beta - 6 areas

Good docs and hardening surfaces exist. Promote after regular upgrade/security scenario runs prove no setup regressions.

Coverage Experimental - 0%Quality Beta - 72%Completeness Beta - 79%Partial - 5

Automation: cron, hooks, tasks, polling - M3 Beta - 6 areas

Documented and usable, but scenario proof should cover unattended delivery, retries, and failure visibility.

Coverage Experimental - 0%Quality Beta - 72%Completeness Beta - 79%None

Media understanding and media generation - M2 Alpha - 6 areas

Broad capability surface exists, but provider variance, file limits, and node/app parity make this not stable yet.

Coverage Experimental - 1%Quality Alpha - 64%Completeness Alpha - 68%None

Voice and realtime talk - M2 Alpha - 6 areas

Multiple implementations exist across Control UI, apps, and providers. Needs latency, failure-mode, and setup scorecards before beta.

Coverage Experimental - 0%Quality Alpha - 61%Completeness Alpha - 68%None

TUI - M2 Alpha - 5 areas

Present in docs and source, but less visible as a primary user workflow. Needs explicit scenario definition.

Coverage Experimental - 0%Quality Alpha - 59%Completeness Alpha - 66%None

ClawHub - M2 Alpha - 4 areas

Public docs and ecosystem concept exist. Needs install, trust, update, rollback, and compatibility scorecards.

Coverage Experimental - 0%Quality Alpha - 58%Completeness Alpha - 62%None

OpenClaw App SDK - M2 Alpha - 6 areas

OpenClaw App SDK is a distinct external app contract separate from Gateway runtime and Plugin SDK. Current scoring shows a real @openclaw/sdk path with gaps around public packaging, auto-discovery, approvals, helpers, and compatibility.

Coverage Experimental - 0%Quality Alpha - 54%Completeness Alpha - 53%None

Platform

Linux Gateway host - M4 Stable - 5 areas

Node runtime is recommended, systemd user service is documented, and VPS/container guidance is broad.

Coverage Experimental - 0%Quality Beta - 75%Completeness Stable - 89%Partial - 4

macOS Gateway host - M4 Stable - 7 areas

LaunchAgent service path, local/remote Gateway modes, CLI install, and app integration are documented.

Coverage Experimental - 0%Quality Beta - 74%Completeness Stable - 88%None

Docker and Podman hosting - M3 Beta - 4 areas

Install docs exist and are common deployment paths. Promote after recurring release smoke captures upgrade and volume behavior.

Coverage Experimental - 5%Quality Beta - 71%Completeness Beta - 79%None

Windows via WSL2 - M3 Beta - 6 areas

Recommended Windows path with systemd/user-service guidance and boot-chain docs. Promote after repeated install/update scorecards.

Coverage Experimental - 3%Quality Alpha - 69%Completeness Beta - 79%Partial - 5

Raspberry Pi and small Linux devices - M3 Beta - 4 areas

Platform docs exist and Gateway path is Linux-based. Needs hardware-specific release smoke proof to move higher.

Coverage Experimental - 0%Quality Alpha - 67%Completeness Beta - 79%None

macOS companion app - M3 Beta - 8 areas

Rich menu bar app, permissions, node mode, Canvas, voice wake, WebChat, and remote mode exist. Still fast-moving enough to avoid Stable.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%None

Android app - M2 Alpha - 7 areas

Public Google Play path exists, but app docs still describe the rebuild as extremely alpha and call out release hardening work.

Coverage Experimental - 0%Quality Alpha - 59%Completeness Alpha - 66%None

Native Windows - M2 Alpha - 4 areas

Core CLI/Gateway flows work, but docs still recommend WSL2 for the full experience and list native caveats.

Coverage Experimental - 0%Quality Alpha - 58%Completeness Alpha - 66%Partial - 1

Kubernetes hosting - M2 Alpha - 4 areas

Kubernetes hosting is a distinct Kustomize-based cluster deployment path. Current scoring shows a real minimal deployment path with gaps around Kubernetes-specific CI, ingress/TLS/NetworkPolicy packaging, backup/restore, and production exposure hardening.

Coverage Experimental - 0%Quality Alpha - 55%Completeness Alpha - 61%None

iOS app - M1 Experimental - 8 areas

Internal preview / super-alpha. TestFlight and relay-backed push flows exist, but no public distribution yet.

Coverage Experimental - 0%Quality Experimental - 41%Completeness Experimental - 44%None

Nix install path - M1 Experimental - 5 areas

Optional install flow. Needs clearer support promise before alpha/beta promotion.

Coverage Experimental - 0%Quality Experimental - 41%Completeness Experimental - 44%None

watchOS companion surfaces - M1 Experimental - 5 areas

Source has Watch app/extension surfaces; public docs do not yet present this as a user feature.

Coverage Experimental - 0%Quality Experimental - 41%Completeness Experimental - 44%None

Linux companion app - M0 Planned - 5 areas

Docs say native Linux companion apps are planned; Gateway is the supported Linux path today.

Coverage Experimental - 0%Quality Experimental - 19%Completeness Experimental - 21%None

Native Windows companion app - M0 Planned - 5 areas

Planned only.

Coverage Experimental - 0%Quality Experimental - 19%Completeness Experimental - 21%None

Channel

Discord - M4 Stable - 6 areas

Deep docs and broad feature coverage. Voice/delegation paths should stay separately scored as beta/alpha.

Coverage Experimental - 0%Quality Beta - 73%Completeness Stable - 87%Partial - 4

Telegram - M3 Beta - 5 areas

Core channel is mature enough for regular use, but high-variance UX and media edge cases need recurring scenario proof.

Coverage Experimental - 0%Quality Alpha - 68%Completeness Beta - 78%Full - 5

Slack - M3 Beta - 5 areas

First-class channel docs and routing surface. Needs workspace install/admin scenario scorecards.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%Full - 5

iMessage and BlueBubbles - M3 Beta - 5 areas

Supported iMessage runs through imsg on a signed-in macOS Messages host; legacy BlueBubbles configs require migration. Keep macOS permissions, SSH wrapper, SIP/private API, and migration caveats visible.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%None

WhatsApp - M3 Beta - 5 areas

Core path is important and documented; upstream Baileys/session volatility keeps it below Stable.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%None

Matrix - M2 Alpha - 6 areas

Supported via bundled plugin. Needs bridge, auth, and room lifecycle scorecards.

Coverage Experimental - 0%Quality Alpha - 60%Completeness Alpha - 67%None

Google Chat - M2 Alpha - 5 areas

Documented channel, but enterprise/admin setup raises maturity risk.

Coverage Experimental - 0%Quality Alpha - 59%Completeness Alpha - 66%None

Microsoft Teams - M2 Alpha - 5 areas

Enterprise auth/admin flows need explicit scenario proof.

Coverage Experimental - 0%Quality Alpha - 59%Completeness Alpha - 66%None

Signal - M2 Alpha - 5 areas

Supported channel docs exist; needs stronger install and reconnect proof.

Coverage Experimental - 0%Quality Alpha - 59%Completeness Alpha - 66%None

Feishu, QQ Bot, WeChat, Yuanbao, Zalo, Zalo Personal, regional channels - M2 Alpha - 4 areas

Important regional coverage, but public support level should be calibrated per account type, upstream approval, and maintainer proof.

Coverage Experimental - 0%Quality Alpha - 55%Completeness Alpha - 58%None

Mattermost, LINE, IRC, Nextcloud Talk, Nostr, Twitch, Tlon, Synology Chat - M2 Alpha - 4 areas

Supported surfaces exist, but maturity likely varies by upstream and maintainer coverage. Score individually later.

Coverage Experimental - 0%Quality Alpha - 53%Completeness Alpha - 54%None

Voice Call channel - M1 Experimental - 5 areas

Optional/plugin path with complex realtime behavior. Needs scenario scorecard before public beta.

Coverage Experimental - 0%Quality Experimental - 41%Completeness Experimental - 44%None

Provider and tool

Browser automation, exec, and sandbox tools - M3 Beta - 3 areas

Core tools are documented, but host security and permission UX should stay under active scorecard review.

Coverage Experimental - 15%Quality Beta - 75%Completeness Beta - 79%Partial - 2

OpenAI and Codex provider path - M3 Beta - 5 areas

Deep docs, OAuth/subscription path, realtime voice, image, and compatibility behavior. Provider churn keeps this from Stable without release-scorecard proof.

Coverage Experimental - 8%Quality Beta - 74%Completeness Beta - 79%Partial - 3

Web search tools - M3 Beta - 4 areas

Multiple providers and docs exist. Needs quota/error/SSRF proof per provider family.

Coverage Experimental - 7%Quality Beta - 74%Completeness Beta - 79%None

Anthropic provider path - M3 Beta - 5 areas

First-class model provider. Needs recurring auth/catalog/tool-call scenario proof.

Coverage Experimental - 0%Quality Beta - 71%Completeness Beta - 78%None

Google provider path - M3 Beta - 5 areas

First-class provider with model and realtime surfaces. Needs separate Live/Talk scoring.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%None

OpenRouter provider path - M3 Beta - 4 areas

Unified provider path is documented and valuable, but model-specific behavior varies.

Coverage Experimental - 0%Quality Alpha - 66%Completeness Beta - 78%None

Image, video, and music generation tools - M2 Alpha - 5 areas

Capability exists across providers, but quality, latency, and parameter compatibility vary too much for beta without per-provider proof.

Coverage Experimental - 0%Quality Alpha - 61%Completeness Alpha - 68%None

Local model providers: Ollama, vLLM, SGLang, LM Studio - M2 Alpha - 5 areas

Useful and documented, but environment variance is high.

Coverage Experimental - 0%Quality Alpha - 61%Completeness Alpha - 68%None

Long-tail hosted providers - M2 Alpha - 3 areas

Many docs/reference pages exist; score should be generated from provider metadata plus live smoke coverage.

Coverage Experimental - 0%Quality Alpha - 61%Completeness Alpha - 68%None

​Maturity taxonomy

​How to read this page

​Maturity levels

​Product areas

​Details

​Core

​Platform

​Channel

​Provider and tool

Maturity taxonomy

How to read this page

Maturity levels

Product areas

Details

Core

Platform

Channel

Provider and tool