Batch Web Research With Parallel AI

I want to batch Research, Run in Parallel.

INSTALL (run in my terminal)
  curl -fsSL https://composio.dev/install | bash
  composio login

You have access to 1000+ app integrations through the Composio CLI.
search → find tools. execute → run them. link → connect accounts.

Workflow: search → execute. If execute fails with an auth error, run link, then retry. Don't preemptively link accounts. Bias toward action — input validation and auth errors are self-descriptive, just try it.

CORE COMMANDS

  search
    Find tools. Use this first — describe what you need in natural language.
    composio search <query> [--toolkits <slug>] [--limit <n>]

  execute
    Run a tool. Handles input validation and auth checks automatically.
    If auth is missing, the error tells you what to run.
    composio execute <slug> -d '<json>'

  link
    Connect an account. Only when execute tells you to.
    composio link <toolkit>

EXAMPLE

  composio search "create github issue"
  # → returns GITHUB_CREATE_ISSUE
  composio execute GITHUB_CREATE_ISSUE -d '{ repo: "owner/repo", title: "Bug" }'
  # → if auth missing: "Run `composio link github` first"
  composio link github
  # then retry execute

TASK: Batch Web Research With Parallel AI
Likely toolkits: Parallel AI.

Steps:
  1. Create a Parallel AI task group for the batch
  2. For each topic, start a research run with your structured output schema
  3. Poll the task group until every task hits a terminal state
  4. Fetch each completed task's structured output
  5. Aggregate into a Google Doc or Sheet if an output target is set
  6. Return the aggregated link and per-task status

Ask me before any action that sends messages or modifies data in a connected app.