Skip to main content
Glama
Sign In
SuperDev

SuperDev

by Xsxdot
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Go
Local

See it in action: step through a real troubleshooting run — from the user's symptom, to live logs, to a breakpoint on the already-running process, to an approved restart, to browser verification: gosuper.dev/#troubleshooting.

Why SuperDev

AI coding tools can read, edit, and run code. But that only answers "what's in the repo." Real development happens in the runtime: which services are up, which ports are taken, which logs belong to this feature, where a bug actually is, and whether the frontend you just changed even renders.

A human developer has a full toolbox for this — a process list, a log viewer, a debugger that stops on a line, a browser to click through the UI. Your AI has almost none of it. It writes the code, then goes blind: it can't see the running service, can't step through the bug, can't tell if the page broke. So it guesses, restarts services it can't see, and asks you to paste logs.

SuperDev gives AI the same workbench you use — See, Inspect, Operate — over one source of runtime truth that the desktop app and AI share, across every local and remote project.

Related MCP server: Daemoon

See, Inspect, Operate

A human developer doesn't just read code — they watch it run, stop it mid-flight to inspect state, and drive the UI to confirm it works. SuperDev gives AI all three.

👁 See — runtime and logs

AI sees the services already running (no shadow copies, no port contention), and reads the same live/historical logs you do — cross-service search, context lookup, folding, bookmarks — instead of asking you to paste fragments. Diagnostics give deterministic evidence; AI owns the root-cause reasoning.

🔬 Inspect — breakpoint debugging

When logs aren't enough, AI attaches to a running managed process (no restart, same pid), stops on a source line, and reads the call stack, scopes, and variables in one call. Default support covers Go, Python, Rust, C/C++; experimental support covers Node and Java/Kotlin (bring your own / configured adapter). Exposed over MCP as list_code_debug_targets and debug_capture_at.

🎮 Operate — browser and deploys

AI changes the frontend, then verifies it itself: navigate, click, type, screenshot, read console logs and network requests, and (when authorized) evaluate JS in the page — driven through Playwright. For shipping, AI runs the same DAG pipelines you do. Every real-environment action stays behind preflight, approval, one-time tokens, and audit.

Highlights

See / Inspect / Operate is what AI does. Two things make it trustworthy and continuous — one shared source of runtime truth and safe operations. Everything below serves those.

🤝 Shared runtime, no competing services

  • AI observes existing services, ports, logs, and deployments before deciding whether it needs to request an action.

  • Avoid shadow environments, port contention, duplicate processes, and split runtime state.

  • Follow one feature across local services, log changes, pipeline runs, remote deployments, and ingress.

  • Treat production errors as shared runtime context, not pasted text detached from the system that produced it.

🔒 Safe AI operations

  • superdev-mcp ships with the desktop app and connects to the local agent at http://127.0.0.1:57017 by default.

  • The bundled SuperDev skill teaches AI to build a global view first, collect evidence, reason explicitly, and only then execute safely.

  • Runtime writes call start/stop/restart directly; when approval is required, MCP waits for desktop approval by default and resumes with a one-time token.

  • Approval tokens are bound to an operation fingerprint, expire quickly, are single-use, and cannot be reused for a different target.

  • Approvals, rejections, executions, and failures are recorded locally for audit.

  • Browser control (navigate / click / evaluate) runs under the same model: open requests approval, evaluate is gated by a trust toggle rather than per-call prompts, and control actions are redacted in the audit log.

Capabilities built around those two

Capability

What it does

Unified runtime console

See local processes, Launchd jobs, systemd services, Docker containers, and remote-host deployments in one model; choose managed control or monitor-only; desktop UI and MCP share one source of truth.

Logs & diagnostics

Live / historical logs, cross-service search, context lookup, filter rules, split panels, synchronized recording, bookmark ranges, repeated-log folding. Diagnostics give deterministic evidence; AI owns the root-cause reasoning.

AI breakpoint debugging

When logs aren't enough, AI attaches to a running managed process (no restart, same pid), stops on a source line, and reads call stack / scopes / variables in one call. Default support: Go, Python, Rust, C/C++. Experimental (bring/configure adapter): Node, Java/Kotlin. Exposed over MCP as list_code_debug_targets / debug_capture_at.

Browser control

AI drives the running frontend through Playwright: browser_navigate / browser_click / browser_type / browser_screenshot / browser_set_viewport / browser_console_logs / browser_network_requests / browser_evaluate, plus snapshot, reload, wait-for-selector, press-key, select-option. Verifies its own UI changes; control actions stay under approval + redacted audit.

Production-minded pipelines

DAG pipelines, reusable templates, variables, artifacts, run history, replayable run logs; built-in Go / Node / Python / Java / Rust / PHP / Vue+Go templates; systemd uses a release/current layout, rollback reuses the same path.

Declarative ingress

Pipelines deliver artifacts repeatedly; Ingress converges long-lived edge state: domains, DNS, reverse proxy, HTTPS, managed certs. Supports nginx, manual DNS, Cloudflare, Aliyun, ACME DNS-01, and orphan detection.

Zero-touch onboarding

  • Choose any detected built-in Connector on first launch. Eight verified built-ins are available: Claude Code, Codex, Cursor, OpenCode, OpenClaw, Hermes, Kimi Code, and Grok.

    • Full support (Claude Code, Codex, Cursor, Hermes, Grok): automatic MCP + Skill + Session Hook.

    • Standard support (OpenCode, OpenClaw, Kimi Code): automatic MCP + Skill, with a manual Session Hook step.

    • Unknown compatible local Agents can still use Connector/manual materials without a hard-coded whitelist.

    • Pi is deferred from this wave and is not registered as a built-in Connector.

  • Other local MCP-capable Agents can be connected manually with the standard stdio instructions.

  • Cloud or isolated-sandbox Agents cannot reach the local 127.0.0.1 endpoint; they require a future authenticated Remote MCP Gateway and are not presented with a misleading local configuration.

  • Install the MCP connection and the SuperDev guide skill from the desktop app.

  • Seed a local superdev-sample project automatically.

  • Copy one prompt to AI and watch the full loop: inspect logs, request approval, approve in the desktop app, and let MCP continue automatically.

Quick Start

Run from source

git clone https://github.com/Xsxdot/super-dev.git
cd super-dev
cd desktop
pnpm install
pnpm tauri dev

macOS app download instructions will be added after the first public release package is verified. This README intentionally avoids unverified release links.

Try the AI safety demo

  1. Open SuperDev.

  2. Pick a detected Connector during onboarding, or open the manual setup instructions for another local MCP Agent.

  3. Install the MCP connection.

  4. Copy the generated prompt into your AI coding agent.

  5. When AI asks to restart the sample service, approve it in SuperDev Operation Approvals.

  6. MCP fetches the one-time approval token and continues the restart; AI reads logs again and explains the WARN/ERROR lines.

Architecture

flowchart TB
    AI["Built-in or manually configured local Agent"] --> MCP["superdev-mcp"]
    MCP --> Agent["Local SuperDev Agent (Go)"]
    Desktop["Desktop UI (Tauri + Vue)"] --> Agent
    Agent --> Runtime["Runtime Control"]
    Agent --> Logs["Logs & Diagnostics"]
    Agent --> Pipelines["Pipelines & Artifacts"]
    Agent --> Ingress["Ingress / DNS / HTTPS"]

The important boundary is the local agent. It is the runtime gateway and source of truth. MCP does not bypass the agent to edit config files, SQLite, processes, or remote hosts. The safety gate is enforced in the agent layer, not merely suggested by prompts.

Examples and Templates

examples/ contains small projects used to validate built-in pipeline templates:

Example

Template

Runtime

go-http

go-binary-build

Go binary

node-http

node-standard-build

Node

python-http

python-standard-build

Python

java-springboot

java-maven-build

Java Spring Boot

rust-http

rust-cargo-build

Rust binary

php-http

php-standard-build

PHP built-in server

vue-go-combined

vue-go-combined-build

Go serving Vue dist

mcp-log-lab

runtime/log diagnostics fixture

Go command services

Ingress examples live in examples/ingress/ and cover manual DNS, Cloudflare, Aliyun, nginx, and TLS declarations.

Status and Roadmap

SuperDev is approaching its first open-source release. The current focus is macOS desktop usage and local-first workflows.

  • Available: Tauri desktop app, Go local agent, MCP server, SuperDev skill, multi-service logs, breakpoint debugging (Go/Python/Rust/C++ by default; Node and Java/Kotlin experimental), browser control (Playwright-driven), operation approvals, pipeline templates, ingress, and zero-touch onboarding.

  • Near term: verified release packaging, final README screenshots, more pipeline templates, a smoother remote agent / tunnel experience, and a richer release walkthrough.

  • Principle: local-first by default. AI can participate in operations, but writes must remain preflighted, approved, token-bound, and auditable.

Platform Support

Platform

Local desktop app

Remote agent install target

Go/Python/Rust/C/C++ debugging

Node debugging

JVM debugging

macOS (darwin)

Supported with Tauri sidecars.

superdev-agent-darwin-amd64 and superdev-agent-darwin-arm64.

Supported by the default attach flow when the language debugger is installed.

Experimental attach via SIGUSR1 inspector activation.

Experimental; bring or configure the JVM adapter.

Linux

Covered by desktop-linux CI packaging.

superdev-agent-linux-amd64 and superdev-agent-linux-arm64.

Supported by the default attach flow when the language debugger is installed.

Experimental attach via SIGUSR1 inspector activation.

Experimental; bring or configure the JVM adapter.

Windows

Covered by desktop-windows CI packaging with .exe sidecars.

superdev-agent-windows-amd64.exe.

Supported by the default attach flow when the matching debugger toolchain is available.

Experimental prearm flow using --inspect=0 instead of Unix signals.

Experimental; bring or configure the JVM adapter.

Development

# Desktop
cd desktop
pnpm install --frozen-lockfile
pnpm build          # Frontend type-check + vite build (does not package the desktop app)
pnpm test           # Frontend unit tests (vitest)
pnpm tauri build    # Package the macOS desktop app (produces an installer, does not launch it)

# Agent
cd agent
go test ./...

pnpm build only builds frontend assets. To package the full macOS desktop app use pnpm tauri build; to just run it and see it live use pnpm tauri dev.

Tauri builds package sidecar binaries through desktop/scripts/build-agent.sh: superdev-agent, superdev-mcp, and superdev-sample.

Open Source Governance

Contributing

Contributions are welcome: pipeline templates, runtime adapters, log diagnostics, ingress providers, documentation, and example projects. Please include reproducible verification commands in PRs when possible. For changes that touch runtime writes, preserve the preview, approval, and audit boundaries.

License

See LICENSE.

Available Tools

79 tools
acquire_test_databaseAcquire test databaseA

Acquire an isolated real test environment (a PostgreSQL database cloned from the project's dev database, plus a free Redis db) and return plaintext connection strings. Use this whenever tests need a real database — never fall back to sqlite or an in-memory substitute. The PG database is a clone of the project's dev database, so it already has schema and seed data. Auto-reclaimed at expiry; call release_test_database when done. Requires project and purpose.

Inputs: purpose (required), project_id (required), kinds, ttl_minutes, +2 more.

analyze_trace_logsAnalyze trace logsC
Read-only

Collect deterministic trace/request log evidence without claiming root cause.

Inputs: limit, after_ms, trace_id, before_ms, +3 more.

append_debug_session_noteAppend debug session noteB

Append a local diagnostic note or observation only; it does not change runtime state or configuration.

Inputs: type (required), actor (required), summary (required), session_id (required), +1 more.

append_log_analysis_to_sessionAppend log analysis to debug sessionA

Run deterministic log analysis and append the result as a local diagnostic record only; it does not change runtime state or configuration.

Inputs: session_id (required), analysis_type (required), to, from, +9 more.

apply_config_changeApply config changeC

Apply a config upsert through the local agent safe-operation flow.

Inputs: kind (required), project, service, pipeline, +5 more.

browser_clickBrowser clickC

Click an element in a browser debug session page.

Inputs: selector (required), session_id (required).

browser_console_logsBrowser console logsB
Read-only

Read recent console logs captured from a browser debug session page.

Inputs: session_id (required), level, limit.

browser_evaluateBrowser evaluateC

Evaluate JavaScript in a browser debug session page and return the serializable result.

Inputs: expression (required), session_id (required).

browser_navigateBrowser navigateA

Perform same-origin full page navigation in a browser debug session. This may reload the page and lose SPA in-memory state; prefer browser_click for SPA route changes.

Inputs: session_id (required), url, path, wait_until.

browser_network_requestsBrowser network requestsA
Read-only

Read recent network request summaries captured from a browser debug session page.

Inputs: session_id (required), limit, filter.

browser_press_keyBrowser press keyA

Press a keyboard key in a browser debug session, optionally focusing a selector first.

Inputs: key (required), session_id (required), selector.

browser_reloadBrowser reloadC

Reload the current page in a browser debug session.

Inputs: session_id (required), wait_until.

browser_screenshotBrowser screenshotC
Read-only

Capture a PNG screenshot from a browser debug session page.

Inputs: session_id (required), full_page.

browser_select_optionBrowser select optionC

Select an option by value or label in a browser debug session page.

Inputs: selector (required), session_id (required), label, value.

browser_set_viewportBrowser set viewportA

Set the viewport size for a browser debug session page before snapshot or screenshot checks.

Inputs: width (required), height (required), session_id (required).

browser_snapshotBrowser snapshotC
Read-only

Read a text snapshot from a browser debug session page.

Inputs: session_id (required), max_text, selector, max_elements.

browser_typeBrowser typeC

Type or fill text into an element in a browser debug session page.

Inputs: text (required), selector (required), session_id (required), fill.

browser_wait_for_selectorBrowser wait for selectorB
Read-only

Wait until a selector reaches a requested state in a browser debug session.

Inputs: selector (required), session_id (required), state, timeout_ms.

close_browser_debug_sessionClose browser debug sessionC
Destructive

Close a browser debug session created by SuperDev.

Inputs: session_id (required).

close_debug_sessionClose debug sessionB

Close a local diagnostic session record only; it does not change runtime state or configuration.

Inputs: session_id (required), summary.

create_debug_sessionCreate debug sessionA

Create a local diagnostic session record only; it does not change runtime state or configuration.

Inputs: title (required), question (required), env_name, project_id, +4 more.

debug_capture_atDebug capture atA

Last-resort code debug for a deployment: stop at a source line and return stack/scopes/variables in one call. Use logs and diagnose tools first.

Inputs: line (required), source (required), deployment_id (required), thread_id, +5 more.

debug_continueDebug continueB

Low-level DAP escape hatch: continue one paused thread for a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_evaluateDebug evaluateB

Low-level DAP escape hatch: evaluate an expression inside the debuggee. This is approval-gated and expression-level audited.

Inputs: expression (required), deployment_id (required), frame_id, approval_token, +1 more.

debug_inspectDebug inspectA
Read-only

Inspect a paused deployment debug runtime and return stack/scopes/variables in one call. Prefer this over chaining low-level DAP tools.

Inputs: deployment_id (required), frame_id, thread_id, max_variables, +1 more.

debug_pauseDebug pauseC

Low-level DAP escape hatch: pause one thread for a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_scopesDebug scopesB
Read-only

Low-level DAP escape hatch: read scopes for one stack frame.

Inputs: frame_id (required), deployment_id (required).

debug_stack_traceDebug stack traceB
Read-only

Low-level DAP escape hatch: read stack frames from a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_step_inDebug step inA

Low-level DAP escape hatch: step into one paused thread in a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_step_outDebug step outB

Low-level DAP escape hatch: step out from one paused thread in a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_step_overDebug step overC

Low-level DAP escape hatch: step over one paused thread in a deployment debug runtime.

Inputs: thread_id (required), deployment_id (required).

debug_variablesDebug variablesA
Read-only

Low-level DAP escape hatch: read variables for one scope. Secret-looking values are redacted by the agent.

Inputs: deployment_id (required), variables_reference (required).

deploy_project_pipelineDeploy project pipelineC

Execute a project-level pipeline deploy or rollback through the local agent.

Inputs: env_name (required), pipeline_id (required), host_ids, variables, +6 more.

describe_language_runtime_schemaDescribe language runtime schemaA
Read-only

Return the config field schema for a language runtime provider, such as go. Use before creating a service so you fill cwd/env/config fields instead of guessing a command string.

Inputs: language (required).

diagnose_serviceDiagnose serviceA
Read-only

Collect runtime status and recent log evidence for one deployment without claiming root cause.

Inputs: env_name, project_id, service_id, project_name, +5 more.

follow_logsFollow logsA
Read-only

Poll recent deployment logs for a bounded duration; use for short tail -f style observation.

Inputs: level, limit, env_name, project_id, +10 more.

get_debug_credentialsGet debug credentialsA
Read-only

Return plaintext debug credentials (test login/password, service api-key) for AI to log in or authenticate legitimately during debugging. Call this when has_debug_credentials/debug_credential_hints are visible and authenticated API testing is needed. Use auth_hint and credential desc to choose the right credential; do not fabricate tokens or bypass auth. project_id|project_name required; pass service to merge project+service level (service overrides). Read-only, not approval-gated.

Inputs: project_id, service_id, project_name, service_name.

get_debug_sessionGet debug sessionA
Read-only

Read one local diagnostic session and its events.

Inputs: session_id (required), limit.

get_log_contextGet log contextA
Read-only

Fetch log context around one log entry by project or deployment.

Inputs: id (required), limit, after_ms, before_ms, +3 more.

get_operation_approvalGet operation approvalA
Read-only

Read one operation approval and return a one-time token when approved.

Inputs: approval_id (required).

get_projectGet projectA
Read-only

Return one SuperDev project by ID or name.

Inputs: project_id, project_name.

get_project_configGet project configC
Read-only

Read an editable project config snapshot from the local agent.

Inputs: root_path, project_id.

get_runtime_snapshotGet runtime snapshotA
Read-only

Return a SuperDev-wide runtime snapshot for projects, services, and deployments.

import_pipeline_templateImport pipeline templateB

Import a pipeline template file into the local agent template library only.

Inputs: path (required), approval_token, debug_session_id.

list_browser_targetsList browser targetsA
Read-only

Return local frontend deployments configured with web entrypoints.

list_code_debug_targetsList code debug targetsA
Read-only

List local managed language runtime deployments that can use last-resort code debugging. Node and JVM (Java/Kotlin) targets are experimental.

list_debug_browsersList debug browsersA
Read-only

Return local Chromium-compatible browsers configured for SuperDev frontend debugging.

list_debug_sessionsList debug sessionsB
Read-only

List local diagnostic session records from the SuperDev agent.

Inputs: limit, status, project_id, project_name.

list_hostsList hostsA
Read-only

Return host selection records. Use non-self host id fields as canonical remote host_ids values; name is display-only.

list_language_runtime_providersList language runtime providersA
Read-only

Return languages with runtime providers. Use before creating a language service so you can choose a supported provider.

list_operation_approvalsList operation approvalsB
Read-only

List pending or historical operation approval requests.

Inputs: limit, status, project_id, project_name.

list_operation_auditList operation auditC
Read-only

List local operation safety audit events.

Inputs: kind, limit, since, project_id, +2 more.

list_pipeline_artifactsList pipeline artifactsB
Read-only

List project-level pipeline artifact history.

Inputs: pipeline_id (required), project_id, project_name.

list_pipeline_runsList pipeline runsC
Read-only

List project-level pipeline execution history.

Inputs: pipeline_id (required), project_id, project_name.

list_projectsList projectsA
Read-only

Return SuperDev projects registered in the local agent.

list_servicesList servicesB
Read-only

Return services and deployment runtime state from the local SuperDev agent.

Inputs: project_id, project_name.

list_test_databasesList test databasesA
Read-only

List active isolated real test environments without connection strings.

open_browser_debug_sessionOpen browser debug sessionB

Open a local frontend deployment in an isolated debug browser and return CDP WebSocket endpoints.

Inputs: deployment_id (required), path, browser_id, open_devtools, +5 more.

preview_config_changePreview config changeB
Read-only

Preview a project/service/project-pipeline config upsert without writing YAML.

Inputs: kind (required), project, service, pipeline, +5 more.

preview_operationPreview operationA
Read-only

Create a deterministic safety preflight plan for one write operation without creating an approval request.

Inputs: kind (required), path, env_name, project_id, +8 more.

preview_pipeline_templatePreview pipeline templateA
Read-only

Dry-run parse and validate a pipeline template YAML string or file.

Inputs: path, yaml.

preview_service_executionPreview service executionA
Read-only

Preview the execution plan for a language service intent without starting it. Use this to explain what start_dev/start_normal/debug_launch would run.

Inputs: language (required), project_root (required), cwd, env, +3 more.

probe_project_configProbe project configB
Read-only

Probe a project directory through the local agent without writing config.

Inputs: root_path, project_id.

read_pipeline_run_logsRead pipeline run logsC
Read-only

Read stored logs for one project-level pipeline run.

Inputs: run_id (required), pipeline_id (required), limit, before, +4 more.

release_test_databaseRelease test databaseA

Release an isolated real test environment before its lease expires.

Inputs: lease_id (required).

renew_test_databaseRenew test databaseB

Renew an isolated real test environment lease before it expires.

Inputs: lease_id (required), ttl_minutes.

restart_serviceRestart serviceA
Destructive

Restart one resolved deployment. If approval is required, wait for desktop approval by default and resume with a one-time token.

Inputs: env_name, project_id, service_id, project_name, +5 more.

search_logsSearch logsC
Read-only

Search SuperDev logs by project or deployment.

Inputs: q (required), limit, cursor_id, project_id, +3 more.

set_debug_breakpointsSet debug breakpointsB

Low-level DAP escape hatch: set source breakpoints for a deployment debug runtime.

Inputs: lines (required), source (required), deployment_id (required).

start_serviceStart serviceA

Start one resolved deployment. If approval is required, wait for desktop approval by default and resume with a one-time token.

Inputs: env_name, project_id, service_id, project_name, +5 more.

stop_serviceStop serviceC
Destructive

Stop one resolved deployment. If approval is required, wait for desktop approval by default and resume with a one-time token.

Inputs: env_name, project_id, service_id, project_name, +5 more.

suggest_service_runtimeSuggest service runtimeA
Read-only

Suggest schema-shaped runtime config for a language service from project_root and cwd. Use the result as a draft, then validate it.

Inputs: language (required), project_root (required), cwd.

summarize_error_windowSummarize error windowA
Read-only

Summarize deterministic error signals in a project or deployment time window without claiming root cause.

Inputs: to, from, limit, since, +3 more.

tail_logsTail logsA
Read-only

Fetch a finite recent deployment log page and optionally scan older pages for level/since filters.

Inputs: level, limit, since, before, +11 more.

upsert_project_configUpsert project configD

Create or edit project base config through the local agent.

Inputs: kind (required), project, service, pipeline, +5 more.

upsert_project_pipelineUpsert project pipelineC

Create or edit one project-level pipeline through the local agent.

Inputs: pipeline (required), project, service, root_path, +4 more.

upsert_serviceUpsert serviceB

Create or edit one service and its deployments through the local agent. Set service.language when using local managed runtime.type=language.

Inputs: service (required), project, pipeline, root_path, +4 more.

validate_project_pipelineValidate project pipelineA
Read-only

Validate an already saved project-level pipeline without executing it.

Inputs: env_name (required), pipeline_id (required), variables, project_id, +2 more.

validate_service_runtimeValidate service runtimeA
Read-only

Validate cwd/env/config for a language service and return diagnostics before upsert_service.

Inputs: language (required), project_root (required), cwd, env, +1 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    C2.9/5.0

    Scored across 79 tools

    Disambiguation3/5

    Most tools target distinct resources or actions, and descriptions are generally detailed. However, there is meaningful overlap among log/diagnostic tools (follow_logs vs tail_logs, analyze_trace_logs vs summarize_error_window) and among config-edit/preview paths (upsert_project_config vs apply_config_change vs preview_config_change), which could cause an agent to pick the wrong one.

    Naming Consistency5/5

    Tool names follow a very consistent snake_case verb_noun or domain_verb pattern (browser_*, debug_*, list_*, upsert_*, preview_*). Even with 79 tools, the naming scheme is predictable and makes categories easy to identify.

    Tool Count1/5

    79 tools is an extreme mismatch for an MCP server surface. Even though the server covers many subdomains, the sheer count makes tool selection and discoverability very difficult for agents.

    Completeness4/5

    The tool set covers broad end-to-end workflows: projects, services, deployments, pipelines, browser debugging, code debugging, logs, test databases, and approvals. Minor gaps exist, such as no destructive delete operations for projects/services/pipelines and no advanced browser interactions beyond core actions.

    Maintenance

    ActivityMaintained
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • Daemoon

      daemoon-dev
      A
      license
      Not graded
      quality
      D
      maintenance
      Gives AI coding agents (Claude Code, Cursor, etc.) unified, secure access to dev infrastructure (Vercel, GitHub, Supabase, Cloudflare, GCP) via a single MCP token.
      MIT

    • AIS Revizor

      bswvladimir-beep
      F
      license
      Not graded
      quality
      D
      maintenance
      A local-first MCP server that gives AI coding agents runtime visibility and AI-managed debug logging. It replaces blind print() debugging by turning runtime execution into causal chains, allowing agents to instantly locate bugs by finding missing .success events in Python and TypeScript code. Single binary with MCP, CLI, and HTTP interfaces.
      -

    • unfour

      zyqzyq
      A
      license
      B
      quality
      B
      maintenance
      Local-first backend developer workspace exposing API debugging, SSH, database, workspace, and diagnostics tools to AI agents through a local MCP server
      72
      2
      Apache 2.0