AVD MCP Server
Provides tools for Android Virtual Device (AVD) automation, including starting emulators, executing ADB commands, and capturing device screenshots.
Allows for the management and automation of Android Virtual Devices (AVD) that have been configured within the Android Studio environment.
Enables the execution of Gradle build commands and tasks as part of automated Android development and testing workflows.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@AVD MCP ServerStart the Pixel_5_API_31 emulator, run pnpm android, and show me a screenshot"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
AVD MCP Server
A Model Context Protocol (MCP) server for Android Virtual Device automation. It can start an emulator, execute commands, and capture screenshots automatically.
Features
Automatically starts Android Virtual Device (AVD) if not running
Executes commands (
pnpm,gradle,npm, etc.)Captures screenshots from the emulator
Returns command output and screenshot in base64 format
Supports
serialselection for device-specific actionsStructured logs and standardized tool errors for better observability
Related MCP server: Android Mobile MCP
Prerequisites
Android SDK with
adbandemulatorin PATHNode.js 18 or higher
At least one AVD configured in Android Studio
Shell available for command execution (
powershellon Windows,shon Linux/macOS, withbashandzshfallback)
Installation
Quick Start
No installation needed. Add this to your MCP client config:
{
"mcpServers": {
"avd-mcp": {
"command": "npx",
"args": ["avd-mcp"]
}
}
}Local Development
git clone https://github.com/jramalho/avd-mcp.git
cd avd-mcp
pnpm install
pnpm build
node dist/index.jsUsage
Project organization for new MCP tools
Keep a consistent pattern for every new tool:
src/
application/
<tool-name>-use-case.ts
mcp/tools/
definitions.ts # inputSchema MCP + zod schemas
handler.ts # dispatcher + error/log wrapper
adapters/node/
<integration>-adapter.ts
command-helpers.ts # runAdb, runEmulator, timeouts
ports/
<integration>-port.ts
shared/
errors/tool-error.ts # friendly + technical error model
logging/logger.ts # structured JSON logs
tests/
mcp-smoke-test.ts # E2E MCP smoke test via stdio client
index.ts # bootstrap do servidor e wiring de dependênciasNaming convention:
Tool name:
avd_<action>(example:avd_start,avd_stop)Use case class:
<Action>AvdUseCaseor<Action>UseCasePort interface:
<Capability>PortAdapter class:
<Runtime><Capability>Adapter(example:AdbAdapter)
Implementation baseline for every tool:
Strong TypeScript typing in input/output contracts and ports
Zod schema validation in
index.tsFriendly error + technical details using
ToolErrorStructured logs via
Logger(tool_call_started,tool_call_succeeded,tool_call_failed)Optional
serialparameter whenever action can target a specific deviceCI-ready options (
noWindow: true, optionalgpuMode: "swiftshader_indirect")
Tool: avd_list
Lists available AVDs from emulator -list-avds.
Tool: avd_start
Starts an AVD with boot options.
Parameters:
avdName(optional)coldBoot(optional): Uses-no-snapshot-load.wipeData(optional): Uses-wipe-data.noWindow(optional): Uses-no-window.readOnly(optional): Uses-read-only.gpuMode(optional): Uses-gpu, allowed values:auto,host,swiftshader_indirect.waitForBoot(optional): Wait for an online device before returning. Defaulttrue.
Tool: avd_stop
Stops an online emulator using adb emu kill.
Parameters:
serial(optional): Emulator serial (example:emulator-5554). If omitted, stops the first online emulator.
Tool: avd_status
Retorna status dos devices conhecidos pelo adb devices em formato estruturado (JSON serializável).
Parameters:
serial(optional): Se informado, retorna status detalhado só desse device.
Output (shape):
requestedSerial(opcional)generatedAtdevices[]com:serial,state,isEmulator,avdName(quando disponível),bootCompletedsummarycom contagens agregadas
Tool: avd_restart
Reinicia um emulador: executa adb emu kill, espera sair do adb devices e inicia novamente reaproveitando a lógica de avd_start.
Parameters:
serial(optional): Emulator serial alvo. Se omitido, usa o primeiro emulador online.coldBoot(optional): Uses-no-snapshot-load.wipeData(optional): Uses-wipe-data.noWindow(optional): Uses-no-window.readOnly(optional): Uses-read-only.gpuMode(optional): Uses-gpu, allowed values:auto,host,swiftshader_indirect.waitForBoot(optional): Aguarda boot completo (sys.boot_completed=1) antes de retornar.
Output (shape):
traceIdtargetSerialavdNamestopDurationMsstartDurationMstotalDurationMsonlineDevicesAfterRestart
Tool: avd_run_and_screenshot
Starts an AVD (if needed), executes a command, waits, and captures a screenshot.
Parameters:
avdName(optional): AVD name. If omitted and no device is online, the first available AVD fromemulator -list-avdsis used.serial(optional): Target online device serial (example:emulator-5554). If provided, no auto-start is attempted.command(required): Command to execute.coldBoot(optional): Uses-no-snapshot-load.wipeData(optional): Uses-wipe-data.noWindow(optional): Uses-no-window.readOnly(optional): Uses-read-only.gpuMode(optional): Uses-gpu, allowed values:auto,host,swiftshader_indirect.waitMsAfterRun(optional): Wait time before screenshot. Default2000.
Example:
{
"avdName": "Pixel_5_API_31",
"serial": "emulator-5554",
"coldBoot": true,
"noWindow": false,
"readOnly": false,
"gpuMode": "host",
"command": "pnpm android",
"waitMsAfterRun": 5000
}Tool: adb_install_apk
Instala APK no device via adb install -r.
Parameters:
serial(optional): Serial alvo.apkPath(required): Caminho local do APK.timeoutMs(optional): Timeout da execução.
Tool: adb_uninstall
Remove pacote do device via adb uninstall.
Parameters:
serial(optional): Serial alvo.packageName(required): Exemplocom.example.app.timeoutMs(optional): Timeout da execução.
Tool: adb_shell
Executa adb shell <command> em safe mode.
Parameters:
serial(optional): Serial alvo.command(required): Comando shell único.timeoutMs(optional): Timeout da execução.
Tool: adb_logcat
Lê logcat com timeout curto e limite de linhas.
Parameters:
serial(optional): Serial alvo.filter(optional): Filtro de prioridade/tag.maxLines(optional): Quantidade máxima de linhas retornadas.timeoutMs(optional): Timeout da execução.
Tool: screenrecord_start
Inicia gravação de tela com adb shell screenrecord e retorna sessionId para finalização posterior.
Parameters:
serial(optional): Serial alvo.maxDurationSeconds(optional): Limite máximo da gravação. Default120.bitRate(optional): Bitrate do vídeo (ex.:4000000).size(optional): Resolução (ex.:1280x720).
Tool: screenrecord_stop
Finaliza gravação por sessionId, faz pull do arquivo para artifacts e remove o arquivo temporário do device.
Parameters:
sessionId(required): Session retornada porscreenrecord_start.serial(optional): Validado contra a sessão ativa.inlineBase64(optional): Retorna vídeo em base64 no payload.
Tool: screenshot
Captura screenshot com suporte a crop, compressão e anotação textual.
Parameters:
serial(optional): Serial alvo.crop(optional):{ x, y, width, height }.compressQuality(optional): 0 a 100.annotate(optional): lista de textos{ text, x, y }.inlineBase64(optional): Retorna PNG em base64 no payload.
Tool: network_toggle
Liga/desliga wifi, dados móveis e modo avião.
Parameters:
serial(optional): Serial alvo.wifiEnabled(required): Liga/desliga wifi.dataEnabled(optional): Liga/desliga dados móveis.airplaneMode(optional): Liga/desliga modo avião.
Tool: network_condition
Aplica condição de rede no emulador por perfil pronto (good, slow_3g, lte, offline) ou perfil avançado.
Parameters:
serial(optional): Serial alvo.profile(required):string:
good|slow_3g|lte|offlineobjeto avançado:
{ latencyMs?, packetLoss?, speedKbps? }
Tool: set_location
Define coordenadas GPS no emulador via adb emu geo fix <lon> <lat>.
Parameters:
serial(optional): Serial alvo.latitude(required): entre-90e90.longitude(required): entre-180e180.
Tool: set_battery_state
Define nível e estado de carga da bateria via dumpsys battery.
Parameters:
serial(optional): Serial alvo.level(optional): 0 a 100.charging(optional):truepara carregando,falsepara descarregando.
Tool: set_rotation
Define orientação da tela via settings put system.
Parameters:
serial(optional): Serial alvo.orientation(required):portrait|landscape.
Tool: set_locale
Define locale via setprop e broadcast de locale.
Parameters:
serial(optional): Serial alvo.language(required): ex.pt,en.country(optional): ex.BR,US.
API level limitations
Alguns comandos variam entre versões de Android e builds de sistema (AOSP/OEM). Em especial, svc data, settings put global airplane_mode_on, dumpsys battery set ..., user_rotation e mudanças de locale por setprop podem exigir permissões diferentes, reboot, restart de app ou podem não surtir efeito imediato em APIs mais novas. Para estabilidade em CI, prefira emuladores AOSP com API fixa por pipeline e valide a efetividade no app (não apenas o exit code do adb).
Structured logging
O servidor usa logging estruturado em src/observability/logger.ts com logInfo, logWarn e logError, sempre com payload { traceId, tool, message, data? }.
Formato humano (default): linha legível para terminal.
Formato JSON (uma linha por evento): habilite com
AVD_MCP_JSON_LOGS=true.
Todas as tools MCP registram início e fim da execução com durationMs, incluindo tool, traceId, deviceId e success (true/false) nos logs de conclusão.
Formato padrão de erro das tools
Quando uma tool falha, a resposta textual retorna um JSON padronizado com os campos:
code: código curto da falha.message: mensagem legível para humano.hints(opcional): sugestões de correção.validOptions(opcional): opções válidas para o campo.
Exemplo geral:
{
"code": "INVALID_INPUT",
"message": "Parâmetros inválidos para avd_start.",
"hints": ["Verifique os campos obrigatórios."],
"validOptions": null
}Exemplo: avd_start com avdName inválido
{
"code": "AVD_NOT_FOUND",
"message": "avdName \"Pixel_7_Pro_API_36\" não encontrado.",
"hints": ["Você quis dizer Pixel_7_Pro_API_35?"],
"validOptions": [
"Pixel_7_Pro_API_35",
"Pixel_8_API_34",
"Medium_Tablet_API_34"
]
}Exemplo: avd_start com gpuMode inválido
{
"code": "INVALID_GPU_MODE",
"message": "gpuMode inválido: vulkan.",
"hints": ["Use um dos valores suportados: auto, host, swiftshader_indirect."],
"validOptions": ["auto", "host", "swiftshader_indirect"]
}Exemplo: adb_shell com comando proibido
{
"code": "SHELL_COMMAND_NOT_ALLOWED",
"message": "Comando não permitido em safe mode. Comandos permitidos: pm list packages [filtro], pm grant <package> <permission>, pm clear <package>, am start ..., am force-stop <package>, monkey -p <package> -c android.intent.category.launcher 1, svc wifi enable|disable, svc data enable|disable, settings put|get (global|system) ..., dumpsys battery, dumpsys battery set level|status|plugged <valor>, dumpsys battery reset, getprop [key], setprop persist.sys.locale|language|country <valor>, rm /sdcard/mcp_record_<timestamp>.mp4, input keyevent|tap|swipe|text ...",
"hints": ["Use adb_shell apenas para comandos na allowlist."],
"validOptions": [
"pm list packages [filtro]",
"pm grant <package> <permission>",
"pm clear <package>",
"am start ...",
"am force-stop <package>",
"monkey -p <package> -c android.intent.category.launcher 1",
"svc wifi enable|disable",
"svc data enable|disable",
"settings put|get (global|system) ...",
"dumpsys battery",
"dumpsys battery set level|status|plugged <valor>",
"dumpsys battery reset",
"getprop [key]",
"setprop persist.sys.locale|language|country <valor>",
"rm /sdcard/mcp_record_<timestamp>.mp4",
"input keyevent|tap|swipe|text ..."
]
}Safe mode e allowlist
O servidor valida todos os comandos adb shell em uma função central (src/adb/shell-safety.ts) com allowlist de padrões permitidos. Com isso, comandos fora da lista são bloqueados com erro de segurança.
Com AVD_MCP_SAFE_MODE=true (default), há bloqueios extras para comandos perigosos, incluindo:
rebootereboot bootloaderrmdestrutivo (ex.:rm -rf /, remoções em áreas críticas)formatewipefora de fluxos controlados
Para desenvolvimento local, é possível desativar o modo estrito com AVD_MCP_SAFE_MODE=false. Mesmo assim, a allowlist continua ativa.
Não desative safeMode em CI/prod: isso reduz proteção contra comandos destrutivos e aumenta risco operacional em hosts compartilhados.
Segurança de paths e comandos
Inputs de comando são sanitizados (sem quebra de linha, sem
\0, etc.).Paths locais recebidos por tools são validados para ficar dentro da pasta base.
Pasta base padrão: diretório atual do processo (
process.cwd()), configurável porAVD_MCP_WORKSPACE_DIR.
Tool: get_metrics
Retorna métricas simples desde o start do processo.
Output (shape):
startedAtuptimeMstotalExecutionstools[]com:tool,executions,avgDurationMs
Media configuration
MCP_ARTIFACTS_DIR: diretório raiz de artifacts (default.artifacts).MCP_INLINE_BASE64: padrão global para retorno base64 (true/false, defaultfalse).Estrutura de saída:
${MCP_ARTIFACTS_DIR}/records/<sessionId>.mp4${MCP_ARTIFACTS_DIR}/screenshots/<timestamp>_<traceId>.png
MCP client call examples
From a client perspective, calls follow this shape:
{
"method": "tools/call",
"params": {
"name": "<tool-name>",
"arguments": {}
}
}Example: list AVDs
{
"method": "tools/call",
"params": {
"name": "avd_list",
"arguments": {}
}
}Example: start headless for CI
{
"method": "tools/call",
"params": {
"name": "avd_start",
"arguments": {
"avdName": "Pixel_5_API_31",
"noWindow": true,
"gpuMode": "swiftshader_indirect",
"waitForBoot": true
}
}
}Example: run command + screenshot in a specific serial
{
"method": "tools/call",
"params": {
"name": "avd_run_and_screenshot",
"arguments": {
"serial": "emulator-5554",
"command": "pnpm android",
"waitMsAfterRun": 4000
}
}
}Example: stop a specific emulator
{
"method": "tools/call",
"params": {
"name": "avd_stop",
"arguments": {
"serial": "emulator-5554"
}
}
}Example: get status from all devices
{
"method": "tools/call",
"params": {
"name": "avd_status",
"arguments": {}
}
}Example: get status for one serial
{
"method": "tools/call",
"params": {
"name": "avd_status",
"arguments": {
"serial": "emulator-5554"
}
}
}Example: restart one emulator (headless, wait boot)
{
"method": "tools/call",
"params": {
"name": "avd_restart",
"arguments": {
"serial": "emulator-5554",
"noWindow": true,
"gpuMode": "swiftshader_indirect",
"waitForBoot": true
}
}
}Example: install APK
{
"method": "tools/call",
"params": {
"name": "adb_install_apk",
"arguments": {
"serial": "emulator-5554",
"apkPath": "C:\\builds\\app-debug.apk",
"timeoutMs": 120000
}
}
}Example: uninstall package
{
"method": "tools/call",
"params": {
"name": "adb_uninstall",
"arguments": {
"serial": "emulator-5554",
"packageName": "com.example.app"
}
}
}Example: adb shell (safe mode)
{
"method": "tools/call",
"params": {
"name": "adb_shell",
"arguments": {
"serial": "emulator-5554",
"command": "pm list packages"
}
}
}Example: adb logcat (simple)
{
"method": "tools/call",
"params": {
"name": "adb_logcat",
"arguments": {
"serial": "emulator-5554",
"filter": "*:E",
"maxLines": 100,
"timeoutMs": 4000
}
}
}Example: start screenrecord
{
"method": "tools/call",
"params": {
"name": "screenrecord_start",
"arguments": {
"serial": "emulator-5554",
"maxDurationSeconds": 60,
"bitRate": 4000000,
"size": "1280x720"
}
}
}Example: stop screenrecord
{
"method": "tools/call",
"params": {
"name": "screenrecord_stop",
"arguments": {
"sessionId": "<SESSION_ID>",
"serial": "emulator-5554",
"inlineBase64": false
}
}
}Example: screenshot with crop + annotate
{
"method": "tools/call",
"params": {
"name": "screenshot",
"arguments": {
"serial": "emulator-5554",
"crop": {
"x": 100,
"y": 200,
"width": 900,
"height": 1600
},
"compressQuality": 80,
"annotate": [
{ "text": "Login", "x": 120, "y": 240 },
{ "text": "CTA", "x": 500, "y": 1500 }
],
"inlineBase64": false
}
}
}Example: network slow 3G profile
{
"method": "tools/call",
"params": {
"name": "network_condition",
"arguments": {
"serial": "emulator-5554",
"profile": "slow_3g"
}
}
}Example: toggle airplane mode
{
"method": "tools/call",
"params": {
"name": "network_toggle",
"arguments": {
"serial": "emulator-5554",
"wifiEnabled": false,
"dataEnabled": false,
"airplaneMode": true
}
}
}Example: set Curitiba location
{
"method": "tools/call",
"params": {
"name": "set_location",
"arguments": {
"serial": "emulator-5554",
"latitude": -25.4284,
"longitude": -49.2733
}
}
}Example: set low battery not charging
{
"method": "tools/call",
"params": {
"name": "set_battery_state",
"arguments": {
"serial": "emulator-5554",
"level": 5,
"charging": false
}
}
}Example: set locale pt-BR
{
"method": "tools/call",
"params": {
"name": "set_locale",
"arguments": {
"serial": "emulator-5554",
"language": "pt",
"country": "BR"
}
}
}Automated scenario runners
Para validar fluxos ponta a ponta no MCP, use os runners abaixo:
pnpm test:adb-tools: cenários de install/uninstall/shell/logcat.pnpm test:media-tools: cenários de screenrecord e screenshot (crop/annotate/base64).pnpm test:network-device-tools: cenários de rede, localização, locale e bateria.
Todos os runners geram relatório OK/FAIL no terminal e retornam código de saída 1 quando existe falha em cenário.
Troubleshooting
Emulator not starting
Check available AVDs with
emulator -list-avdsEnsure
emulatoris in PATHEnsure virtualization is enabled (Intel VT-x / AMD-V)
ADB not found
Install Android SDK Platform-Tools
Add platform-tools to PATH (example):
C:\Users\YourUser\AppData\Local\Android\Sdk\platform-tools
License
MIT License - see LICENSE.
Available Tools
1 toolavd_run_and_screenshotC
Sobe AVD se necessário, roda comando (pnpm/gradle/etc) e tira screenshot.
| Name | Required | Description | Default |
|---|---|---|---|
| avdName | No | ||
| command | Yes | ||
| waitMsAfterRun | No |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It mentions starting an AVD, running a command, and taking a screenshot, but lacks details on permissions required, whether the AVD persists after execution, error handling, or what the screenshot output entails (e.g., format, storage location). For a tool with multiple actions and no annotation coverage, this is a significant gap in transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise and front-loaded in a single sentence: 'Sobe AVD se necessário, roda comando (pnpm/gradle/etc) e tira screenshot.' It efficiently outlines the main actions without unnecessary words, though it could be slightly more structured by separating the steps for clarity.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity (multiple actions: start AVD, run command, take screenshot), no annotations, no output schema, and 0% schema description coverage, the description is incomplete. It doesn't address key aspects like what happens if the AVD fails to start, how commands are executed, or what the tool returns (e.g., screenshot data or success status), leaving the agent with insufficient information for reliable use.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, meaning parameters are undocumented in the schema. The description doesn't add any meaning beyond the parameter names: it doesn't explain what 'avdName' refers to, what types of 'command' are valid, or the purpose of 'waitMsAfterRun'. With three parameters and no schema descriptions, the description fails to compensate for this lack of coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Sobe AVD se necessário, roda comando (pnpm/gradle/etc) e tira screenshot.' This translates to 'Starts AVD if necessary, runs a command (pnpm/gradle/etc) and takes a screenshot.' It specifies the verb (start, run, take screenshot) and resource (AVD, command execution). However, since there are no sibling tools, it doesn't need to differentiate from alternatives, so it doesn't achieve the highest score of 5.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives. It doesn't mention prerequisites, scenarios where it's appropriate, or any exclusions. The phrase 'se necessário' (if necessary) implies some conditional usage but doesn't specify what makes it necessary, leaving the agent with minimal context for decision-making.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
1 tool update
v0.1.0- First observed
avd_run_and_screenshot
TDQS
With only one tool, there is no possibility of ambiguity or overlap between tools. The tool's purpose is clearly defined as running commands and taking screenshots in an AVD environment.
Since there is only one tool, naming consistency is inherently perfect. The tool name 'avd_run_and_screenshot' follows a clear verb-based pattern, though with a single tool, consistency across a set cannot be assessed.
One tool is too few for a server named 'AVD MCP Server', which suggests a broader scope related to Android Virtual Device operations. A single tool covering both command execution and screenshot capture feels thin and incomplete for the domain.
The tool surface is severely incomplete for an AVD-focused server. There are obvious gaps such as creating/starting/stopping AVDs, managing device states, installing apps, or performing other common AVD operations, leaving agents unable to handle typical workflows.
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Control real Android and iOS devices with LLM agents — tap, swipe, type, automate flows.
Drive real devices from your AI Coding tool. Embed a client SDK (Unity, Godot, Flutter, iOS/macOS, Android, React Native, Web) in your app, then capture screenshots, traverse the UI tree, inject taps and key events, and run automated test tasks on the physical device over a secure relay.
Drive real Android & iOS devices and web browsers from natural language for mobile + web QA. 290+ tools across device control, app management, automation sessions, browser automation, and flow recording / replay. Bearer-auth — get a token at robotactions.com → Profile → API Tokens.
- LimrunOAuthcom.limrun
Cloud iOS simulators and Android emulators your agent can create, drive, and throw away.
Related MCP Servers
- AlicenseBqualityDmaintenanceEnables AI agents to capture screenshots from Android devices and emulators, and manage connected devices for UI analysis and testing. Supports device listing, screenshot capture, and integrates with Claude Desktop, Gemini CLI, and GitHub Copilot.22324MIT
- FlicenseAqualityCmaintenanceEnables AI agents to interact with Android devices through UI manipulation, screen capture, touch gestures, text input, and app management via ADB. Provides comprehensive mobile automation capabilities including element detection, navigation, and application control for Android device testing and interaction.94-
- AlicenseNot gradedqualityDmaintenanceEnables interaction with Android devices and emulators through ADB, allowing control actions like tapping, text input, screenshots, UI inspection, and app launching through natural language.88ISC
- AlicenseBqualityBmaintenanceEnables AI assistants to interact with Android devices and emulators via ADB, providing tools for screenshots, UI inspection, touch and text input, app management, and device control.427916MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/jramalho/avd-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server