record-demo

por apify

Gravar ou regenerar os GIFs de demonstração do mcpc (o README hero docs/images/mcpc-demo.gif e as fitas focadas em docs/vhs/) com VHS. Use sempre que solicitado a criar,…

npx skills add https://github.com/apify/mcpc --skill record-demo

record-demo: VHS demo GIFs for mcpc

The tapes in docs/vhs/*.tape are VHS scripts that drive a real shell session — VHS types each command, runs it against a live MCP server, captures the terminal, and renders a GIF. The README hero is docs/images/mcpc-demo.gif, built from docs/vhs/mcpc-demo.tape.

This file is the accumulated know-how. Read all of it before touching a tape — most rules below were discovered by hitting the wall, and skipping them wastes whole render cycles (~1–2 min each).

The hero flow (mcpc-demo.tape)

A basic-use story across both transports:

  1. mcpc — empty state (no sessions, no profiles)
  2. mcpc connect mcp.json:filesystem — local stdio server (auto-names @filesystem)
  3. mcpc — session list (now shows the live session)
  4. mcpc @filesystem tools-list
  5. mcpc @filesystem tools-list --json — JSON output, syntax-highlighted, no jq
  6. mcpc connect mcp.apify.com -H "Authorization: Bearer $APIFY_TOKEN" — remote HTTP server (auto-names @apify)
  7. mcpc @apify tools-list
  8. mcpc @apify tools-get search-actors — inspect one tool's input schema
  9. mcpc @apify tools-call search-actors keywords:="web scraper" limit:=3
  10. mcpc @apify close

Ten commands run ~45s; there is no hard 30s cap for this flow.

Style conventions (the current standard — match these)

  • No # comments in the visible script. The commands are self-descriptive.
  • No | head, no 2>/dev/null on visible commands. Show real output even if long — it scrolls naturally. (connect ≈ 86 lines incl. an "Available commands" list, tools-call search-actors ≈ 56, filesystem tools-list --json ≈ 300.)
  • Continuous session — never clear between steps. Put a single blank-line Enter before each command (after the first) so it's separated from the previous output, like a real terminal session.
  • Colored prompt + bold-white typed commands. In the hidden setup block:
    Type 'export PS1="\[\e[1;38;2;25;230;77m\]$\[\e[0m\] \[\e[1;97m\]"'
    Enter
    Type "trap 'tput sgr0' DEBUG"
    Enter
    
    • PS1 is a bold bright-green $, then ends with \[\e[1;97m\] so the typed input renders bold bright-white.
    • The DEBUG trap runs tput sgr0 before every command so the bold-white input does not bleed into command output.
  • Empty state needs a clean home — hidden: Type 'export MCPC_HOME_DIR="$(mktemp -d)"' so mcpc shows "No active MCP sessions / No OAuth profiles".
  • Color is automatic for non-piped commands — mcpc detects the TTY and emits color (256-color, plenty vivid). You only need export FORCE_COLOR=3 (+ COLORTERM=truecolor for exact hex) when a command is piped (mcpc turns color off when stdout isn't a TTY). The current tapes avoid pipes, so they don't need it. (The CLI palette lives in src/cli/output.ts, RAINBOW_SATURATION, bumped to 78% for vividness — that's where the demo colors come from.)

VHS gotchas (these will bite you)

  • ASCII prompt symbols only. Multibyte glyphs (❯, », ▶) break bash prompt rendering under VHS and show up as garbage like 92m]. Use $ (or >), styled with color + bold.
  • Type quoting: use single quotes around any command containing double quotes. A \" inside a double-quoted Type breaks VHS's parser. e.g. Type 'mcpc @apify tools-call search-actors keywords:="web scraper" limit:=3' and Type 'export PS1="\[\e[…m\]$\[\e[0m\] "'.
  • Output/Screenshot paths: must not start with a digit (Output 1-foo.gif fails to parse) and must not be long absolute paths (the parser chokes). Use short, letter-leading, relative names and run vhs from docs/vhs/.
  • Screenshot is unreliable (frequently exits 2 even though the GIF rendered fine). Don't depend on it — pull frames from the finished GIF instead: ffmpeg -y -ss <seconds> -i x.gif -vframes 1 frame.png, then Read the PNG.
  • Renders are slow (~1–2 min each: real-time timeline + Chromium + ffmpeg encode). Render tapes one at a time — a for loop over several blows the 5-minute command timeout. Extracted frames often land mid-typing; sample a few timestamps around when output should be on screen.
  • Hidden connects leak into the recording if bash falls behind. When a tape connects in the hidden setup (so the feature commands run against a ready session, e.g. tools/scripting/grep), a slow connect lets VHS type ahead; bash then echoes the buffered commands and runs the clear after Show, so the setup spills into frame. Pattern that works: type the connect(s), then one generous Sleep (7–8s) so they finish, then clear, then another Sleep (~1.5s) before Show. (Tapes that connect visibly as their first command don't need this — their hidden setup is just fast exports + clear.)

Stdio servers in a headless / proxied box

  • npx-launched stdio servers are too slow here. npx -y <pkg>'s registry round-trip exceeds mcpc's 60s connect handshake, so connect times out (MCP error -32001). Fixes: pre-install the server (npm i -g <pkg>) and put the direct binary in mcp.json (e.g. mcp-server-filesystem, starts in ~0.3s — also a cleaner session header), or use npx --prefer-offline -y <pkg> once the npm cache is warm (~5s; plain npx -y still does the slow registry check even when cached).
  • Puppeteer does NOT work for headless recording. @modelcontextprotocol/server-puppeteer launches Chromium eagerly on startup and hangs/times out as root in the container. Use @modelcontextprotocol/server-filesystem instead (14 recognizable tools, instant). docs/vhs/mcp.json defines the filesystem entry via its global binary — install it first: npm i -g @modelcontextprotocol/server-filesystem.

Auth token (the authenticated step)

  • Always prompt for the token first. Insist on a short-lived, low-permission token from a TEST / throwaway account — never production. Apify: https://console.apify.com/settings/integrations. Tell the user to revoke it as soon as the recording is done.
  • Pass it inline, for the render only: APIFY_TOKEN=… vhs mcpc-demo.tape. The tape references $APIFY_TOKEN (never the literal), typed inside single quotes so bash expands it at run time — the value is never on screen, never in the GIF, never committed. Always verify a connect frame shows $APIFY_TOKEN, not the value.
  • connect auto-names the session: mcp.apify.com → @apify, mcp.json:filesystem → @filesystem. No @name needed.
  • No-token alternative (public, anonymous): mcpc connect "https://mcp.apify.com/?tools=search-actors,fetch-actor-details,docs".

Skills demo server (skills.tape, security.tape)

No public server serves the MCP Skills extension yet, so skills.tape records against the e2e test server (test/e2e/server/index-v2.ts, 2026-07-28, which serves the git-workflow / refunds / daily skill fixtures) under the name mcp.example.com. mcpc forces HTTPS for non-localhost hosts, so front it with TLS:

WITH_SKILLS=true PORT=13456 pnpm exec tsx test/e2e/server/index-v2.ts &
echo "127.0.0.1 mcp.example.com" >> /etc/hosts
openssl req -x509 -newkey rsa:2048 -nodes -keyout key.pem -out cert.pem -days 2 \
  -subj "/CN=mcp.example.com" -addext "subjectAltName=DNS:mcp.example.com"
# then run any small HTTPS → http://127.0.0.1:13456 reverse proxy on 127.0.0.1:443
export NODE_EXTRA_CA_CERTS="$PWD/cert.pem" NO_PROXY="mcp.example.com,$NO_PROXY"
cd docs/vhs && vhs skills.tape

The connect stays hidden: its output names the test server ("e2e-test-server").

Keychain warning (headless only)

On a box with no keyring, the bearer-token connect prints [keychain] OS keychain unavailable, falling back to file-based credential storage …. It is environment-specific (won't appear on a normal desktop with a keyring) and there is no env var to suppress it — it's a logger.warn in src/lib/auth/keychain.ts gated only by keychain availability and JSON mode. A dbus-run-session + gnome-keyring-daemon wrapper does not fix it in this sandbox (raising the dbus fd limit is blocked). Options: leave it (honest), or add a targeted 2>/dev/null to just that one connect command.

Prerequisites

mcpc --version            # the CLI being demoed (npm i -g @apify/mcpc, or build + pnpm link this repo)
vhs --version             # brew install vhs (needs ttyd + ffmpeg on PATH)
mcp-server-filesystem     # npm i -g @modelcontextprotocol/server-filesystem (the stdio demo server)

Headless + root: VHS drives a Chromium (go-rod auto-downloads it to ~/.cache/rod) that refuses to start without --no-sandbox. If renders fail to launch Chromium, wrap the binary once:

CHROME=$(find ~/.cache/rod/browser -name chrome -type f | head -1)
mv "$CHROME" "$CHROME-real"
printf '#!/bin/sh\nexec "$(dirname "$0")/chrome-real" --no-sandbox --disable-gpu --disable-dev-shm-usage "$@"\n' > "$CHROME"
chmod +x "$CHROME"

(Run vhs once first to trigger the Chromium download.)

Render and verify

cd docs/vhs
APIFY_TOKEN=…   vhs mcpc-demo.tape                                   # real-time; be patient
ffprobe -v error -show_entries format=duration -of csv=p=0 mcpc-demo.gif   # check length
ffmpeg -y -ss 12 -i mcpc-demo.gif -vframes 1 /tmp/f.png             # spot-check a frame, then Read it
cp mcpc-demo.gif ../images/mcpc-demo.gif                            # update the README hero

Check each frame class: empty state, stdio connect + session list, tools-list, JSON, remote connect (token NOT visible), tool-call result, close. Then revoke the token.

Optimize the GIF size (do this before committing)

VHS GIFs are large (the hero is ~5 MB raw). Shrink them losslessly-ish with gifsicle (kornel.ski/lossygif) — --lossy drops imperceptible inter-frame detail and cuts ~60% off a text-terminal GIF with no visible quality loss:

# in place; lossy=200 ≈ 60% smaller, text stays crisp (verified by frame diff)
for f in docs/images/mcpc-demo.gif docs/vhs/*.gif; do
  [ "$f" = docs/vhs/mcpc-demo.gif ] && continue   # ignored raw hero
  gifsicle -O3 --lossy=200 -b "$f"
done

After optimizing, re-extract a colored frame (ffmpeg -ss 12 …) and eyeball it — lossy=200 is the sweet spot; going much higher smears the antialiased text. When the optimized hero lands on main, bump the README cache-buster (mcpc-demo.gif?v=N → ?v=N+1): GitHub's camo image proxy caches by full URL, so a new ?v= is what makes it re-fetch the smaller file instead of serving the old cached copy.

What's committed

  • docs/images/mcpc-demo.gif — the README hero (canonical copy).
  • docs/vhs/*.gif — the per-feature recordings are committed too, so they're easy to find and reuse. .gitignore ignores only docs/vhs/mcpc-demo.gif (the hero's raw output, since it's committed under docs/images/).
  • proxy.gif needs a token to record and isn't committed until recorded.

The tapes

TapeRecords
mcpc-demo.tapeHero basic-use flow (stdio + remote) → docs/images/mcpc-demo.gif
quickstart.tapeMinimal connect → list → call
tools.tapetools-list / tools-get / tools-call, inline JSON, stdin
scripting.tape--json piped through jq (code mode)
grep.tapeDynamic tool discovery with mcpc grep across two sessions (Apify + filesystem)
proxy.tapeMCP proxy / AI sandboxing (keeps a bearer token on purpose)
skills.tapeMCP Skills extension: skills-list / skills-get / resources-directory-read against the e2e test server served as mcp.example.com
security.tapeAuto-discovery skipping a checked-in .mcp.json entry that reads ${GITHUB_TOKEN} (fixture: security.mcp.json, copied into a mktemp -d HOME so no real configs are discovered); needs mcp.example.com for the trusted entry

All focused tapes follow the same conventions as the hero (bold $ prompt, bold-white commands, no comments, blank-line separation, mktemp home). quickstart/tools/scripting/grep are token-free (public ?tools= URL); grep also connects a local filesystem stdio server so it can search across two sessions. proxy.tape keeps a bearer token because demonstrating that you can proxy a credentialed session without leaking the token is its entire point (it's the one focused tape that needs a token to record).

Mais skills de apify

apify-influencer-brand-collabs
apify
Descubra parcerias entre marcas e criadores no Instagram encadeando Apify Actors. Use quando o usuário perguntar quem colabora com uma marca, quais marcas um criador fez conteúdo pago…
apify-actor-development
apify
Crie, depure e implante programas serverless em nuvem para web scraping, automação e processamento de dados. Suporta templates em JavaScript, TypeScript e Python com as bibliotecas integradas Crawlee, Playwright e Cheerio para crawling baseado em HTTP e navegador. Inclui testes locais via apify run com armazenamento isolado, validação de esquema para entradas/saídas e implantação na plataforma Apify via apify push. Requer autenticação da CLI Apify e metadados obrigatórios generatedBy em .actor/actor.json para IA...
apify-actorization
apify
Converta projetos existentes em Apify Actors serverless com integração de SDK específica para cada linguagem. Suporta JavaScript/TypeScript (com Actor.init() / Actor.exit()), Python (gerenciador de contexto assíncrono) e qualquer linguagem via wrapper CLI. Fornece fluxo de trabalho estruturado: apify init para estruturar, aplicar encapsulamento do SDK, configurar esquemas de entrada/saída, testar localmente com apify run e depois implantar com apify push. Inclui validação de esquemas de entrada e saída, conteinerização Docker e opcional pay-per-event...
apify-content-analytics
apify
Análise de conteúdo multiplataforma via Apify Actors para Instagram, Facebook, YouTube e TikTok. Suporta mais de 17 Actors especializados cobrindo posts, reels, stories, comentários, hashtags, seguidores e anúncios em todas as quatro plataformas. Busca dinamicamente os esquemas dos Actors usando a CLI mcpc para determinar as entradas necessárias e os campos de saída disponíveis. Gera resultados em três formatos: exibição rápida no chat, exportação CSV ou exportação JSON com contagens de resultados personalizáveis. Requer token Apify no arquivo .env e Node.js 20.6+...
apify-ecommerce
apify
Extraia dados de produtos, preços, avaliações e informações de vendedores de mais de 50 marketplaces de e-commerce. Três modos de fluxo de trabalho: Produtos e Preços (monitoramento de preços, análise de concorrentes), Avaliações de Clientes (análise de sentimento, problemas de qualidade) e Inteligência de Vendedores (descoberta de fornecedores via Google Shopping). Suporta Amazon (mais de 20 regiões), Walmart, eBay, IKEA, Costco e varejistas europeus; entrada por URLs de produtos, URLs de categorias ou pesquisa por palavras-chave. Análise opcional com IA gera insights sobre preços...
apify-generate-output-schema
apify
Gere esquemas de saída (dataset_schema.json, output_schema.json, key_value_store_schema.json) para um Apify Actor analisando seu código-fonte. Use quando…
apify-influencer-discovery
apify
Descubra e avalie influenciadores no Instagram, Facebook, YouTube e TikTok usando Apify Actors. Roteia solicitações de descoberta para mais de 15 Actors especializados, abrangendo scraping de perfis, busca por hashtags, análise de engajamento e descoberta de nichos em todas as principais plataformas. Busca dinamicamente os esquemas dos Actors via mcpc para determinar as entradas necessárias e os campos de saída disponíveis antes da execução. Suporta três modos de exportação: exibição inline no chat, saída em arquivo CSV ou JSON, com contagens de resultados personalizáveis...
apify-ultimate-scraper
apify
Raspador web automatizado que seleciona os melhores Atores para mais de 55 plataformas, incluindo Instagram, TikTok, YouTube, Facebook, Google Maps e outras. Abrange mais de 55 Atores pré-configurados em 8 plataformas principais, com orientação de seleção específica para cada caso de uso (geração de leads, descoberta de influenciadores, monitoramento de marca, análise de concorrentes, pesquisa de tendências). Suporta três formatos de saída: exibição rápida no chat, exportação CSV ou exportação JSON com limites de resultados personalizáveis. Inclui padrões de fluxo de trabalho com múltiplos Atores para cenários complexos...