Short answers. Useful details.
Practical answers about using Pergunta day to day, privacy and control, and API & MCP access.
Pergunta turns a long list of intentions into a small, realistic plan for the day. You capture what is on your mind, and it helps you decide what actually fits — so each day starts with a short, doable plan instead of an overwhelming backlog.
A slipped plan is expected, not a failure. Unfinished items stay in your list, and you re-plan from where you are. Nothing is lost, and you are not asked to catch up on everything at once — you simply decide what fits next.
When your list grows faster than your days, you can scale back to a smaller plan at any time. Pergunta is built around keeping the active plan small, so trimming commitments is a normal, expected step rather than an exception.
A conventional to-do app keeps adding to an ever-growing list. Pergunta focuses on what you can realistically do now: it helps you choose a small daily plan, adjust when the day changes, and review regularly, so the list serves you instead of the other way around.
It is for people who are capable but inconsistent with follow-through — professionals and individuals who plan ambitiously, get pulled off track, and want steady, structured support rather than another tool to maintain.
Yes. Access is authenticated, isolated per account, and protected in transit.
Core service records are stored in the European Union, although some third-party integrations may process limited data outside the European Union where that is needed to operate the service.
No. The access model is designed so each account works only with its own data.
Yes. In Settings, under Privacy & Data, you can export your data or delete your account.
Yes. Certain core product functions may require LLMs to process the inputs you submit and related internal service data flows to generate questions, structure, plans, and other service outputs. Some optional LLM-assisted features, including photo- or vision-based task suggestions, depend on a specific consent, which can be managed in Settings, under Privacy & Data.
When you attach a photo or another file to a task, it is kept as a private attachment scoped to your account. If you choose to use camera capture to suggest a task, image bytes may be sent to a vision-capable LLM processor to generate the suggestion. This feature is optional and depends on the specific consent in Settings, under Privacy & Data.
Voice input and accessibility telemetry are optional features. When active, they depend on explicit consent and can be managed in Settings, under Privacy & Data.
Checkout, recurring billing, customer portal, receipts, and refunds for paid plans are operated by Polar (polar.sh), a third-party payment processor acting as merchant of record. Full payment-instrument data is handled by that processor and is not stored in pergunta.me's own systems. Additional details are in the Privacy Policy and Terms of Use.
API and MCP access is available on the Enthusiast and Pro plans (see /en/pricing/ for details), with a 10-day free trial for your first paid subscription. After access is active, sign in to your account and open Settings → Task API Access → Generate API Key. Your Pergunta key is shown once: store it in a secret manager or environment variable, never publish it, and revoke it and generate a replacement if it is exposed.
In any MCP client with OAuth support, add https://mcp.pergunta.me/mcp and start the connection. The client performs dynamic registration, uses PKCE-compatible authorization, and requests the OIDC scopes openid, profile, and email. Confirm Pergunta consent in the authorization window and let the client return to ChatGPT to finish; you do not create a fixed app, client secret, or ChatGPT token. MCP access requires an active Enthusiast or Pro plan (see /en/pricing/). Verify the connection by running a read-only tool. Use this stable section at /en/faq/#developers if the client asks for the guidance again.
Activate access on an Enthusiast or Pro plan (see /en/pricing/), generate a Pergunta key at Settings → Task API Access → Generate API Key, and copy it when it is shown. Configure the MCP client with https://mcp.pergunta.me/mcp and one of the supported header forms below; keep the key in an environment variable or secret store, never in a repository, screenshot, or public message. Run a read-only call to verify the connection, and revoke/rotate the key in Settings if it is shared or no longer needed. See /en/faq/#developers for this stable guidance.
MCP with API key
export PERGUNTA_API_KEY='YOUR_PERGUNTA_API_KEY'
curl --fail --show-error --silent \
--header "x-api-key: ${PERGUNTA_API_KEY}" \
"https://mcp.pergunta.me/mcp"
# Supported alternative: --header "Authorization: Bearer ${PERGUNTA_API_KEY}"It starts with `tm_`. You can keep multiple active keys and revoke them one by one.
Use your API key as a bearer token.
List inbox tasks
export PERGUNTA_API_KEY='tm_your_api_key'
curl --fail --show-error --silent \
--header "Authorization: Bearer ${PERGUNTA_API_KEY}" \
"https://app.pergunta.me/api/tasks?scope=inbox"Yes. Use the same auth pattern for write requests.
Create a task
export PERGUNTA_API_KEY='tm_your_api_key'
curl --fail --show-error --silent \
--request POST \
--header "Authorization: Bearer ${PERGUNTA_API_KEY}" \
--header "Content-Type: application/json" \
--data '{"title":"Draft weekly plan","status":"inbox","priority":"normal"}' \
"https://app.pergunta.me/api/tasks"Yes. MCP access is available on the Enthusiast and Pro plans (see /en/pricing/ for details), with a 10-day free trial for your first paid subscription. Pergunta MCP uses API-key authentication; pass your key in the Authorization header as a bearer token.
Add the remote HTTP server and pass your API key in the Authorization header.
Claude Code
export PERGUNTA_API_KEY='tm_your_api_key'
claude mcp add --transport http --scope user pergunta https://mcp.pergunta.me/mcp \
--header "Authorization: Bearer ${PERGUNTA_API_KEY}"Register the remote server and read the bearer token from an environment variable.
Codex
export PERGUNTA_API_KEY='tm_your_api_key'
codex mcp add pergunta \
--url https://mcp.pergunta.me/mcp \
--bearer-token-env-var PERGUNTA_API_KEYSet `PERGUNTA_API_KEY` in your environment and place this block in `~/.config/opencode/opencode.json` or your project `opencode.json`.
OpenCode
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"pergunta": {
"type": "remote",
"url": "https://mcp.pergunta.me/mcp",
"enabled": true,
"oauth": false,
"headers": {
"Authorization": "Bearer {env:PERGUNTA_API_KEY}"
}
}
}
}Call the verify endpoint before wiring a full client.
Verify key
export PERGUNTA_API_KEY='tm_your_api_key'
curl --fail --show-error --silent \
--header "Authorization: Bearer ${PERGUNTA_API_KEY}" \
"https://mcp.pergunta.me/mcp/verify"MCP accepts OAuth bearer tokens. MCP uses an API key as a bearer token. For key access, it accepts both `x-api-key: <api-key>` and `Authorization: Bearer <api-key>`. These are Pergunta credentials, not ChatGPT credentials; keep them only in environment variables or secret stores, do not share them, and revoke/rotate them when needed.
Need the full legal details?
Read the Privacy Policy