Skip to content

Frequently asked questions

Short answers. Each links to the detailed explanation.

ChatGPT is a chatbot. Your agent here runs 24/7, reads your email, posts to social media, browses the web, remembers you across sessions, and takes action — not just chats. Full comparison →

MiniMax M2.7 by default, on every tier. If you want to use a different model, add your own API key via BYOK (see below) — your agent will route through your key for any provider you connect, at no token cost to your included usage.

No. Your agent runs in the cloud (in its own private container on our infrastructure). You need internet to reach it. But once connected, it doesn’t need you to stay online — it can keep working while you’re offline.

Yes — BYOK is available today on every tier, including trial. Go to Settings → Agent → Bring Your Own Key and add a key for Anthropic (Claude), OpenAI, Google (Gemini), MiniMax, DeepSeek, OpenRouter, or any OpenAI-compatible endpoint. When a BYOK key is present, your agent routes requests through it directly and we don’t bill any token credits against your included usage — you pay the provider on their bill. A popular combo: buy a MiniMax M2.7 flat-rate coding plan from MiniMax and drop the key in — effectively unlimited model usage for a fixed monthly fee.

Is it really learning about me, or just pretending?

Section titled “Is it really learning about me, or just pretending?”

Really learning. It maintains two memory files (MEMORY.md and USER.md) that are injected into every conversation’s context. Every ~10 turns, it reviews the conversation and updates these files with anything worth remembering. More on memory →

Not directly — one subscription = one agent = one user. If multiple team members need agents, each needs their own account. Team plans with shared workspaces are on the roadmap.

Yes. Each agent runs in an isolated container with its own encrypted storage. Your email credentials, conversations, and memory files never mix with any other user’s data. Full privacy details →

No. Your data (emails, chats, files) is never used to train any model. We don’t license your data to any third parties.

  • You, from your dashboard
  • Your agent, in its own container
  • Nobody else — not even us, except in very specific support cases where you explicitly grant access

Retained for 30 days (so you can come back). After 30 days, permanently deleted. You can export everything before canceling. Canceling details →

Not yet. If you need strict compliance certifications for regulated industries, we’re not the right fit right now.

  • Trial: $10 for 15 days (credited to first month if you subscribe)
  • Standard: $29/mo or $290/yr (2 months free)
  • Pro: $59/mo or $590/yr (2 months free)
  • Email & Calendar: included on every tier — unlimited inboxes, no add-on fee
  • Social post credit packs: $3.40 for 300 extra posts, optional, never expire

Full pricing →

Credits pay for the LLM tokens your agent uses. $5/mo included on Standard, $15/mo on Pro. Roughly 250 tasks/month at Standard. More on credits →

Yes. No contracts, no hoops. The cancel flow lives at the very bottom of Dashboard → Settings → Billing as a small muted link — we didn’t want a big red button sitting above everything you actually come here to do. Cancellation is immediate (not period-end) and you get a partial refund: $5 base + the dollar value of any unused credits − Stripe processing fees. Billing errors still get full refunds via support@. Canceling →

Can I pause my agent instead of canceling?

Section titled “Can I pause my agent instead of canceling?”

Yes — that’s a real feature. Settings → Agent → Agent Trust & Approvals → Pause automation lets you pause for 1h / 4h / 24h / 7 days. While paused, every auto action requires your approval instead, so your agent stops acting on its own but your workspace, memories, and subscription stay intact. Lower commitment than canceling and the quickest fix if you just need a quiet window. Trust levels →

Monthly included credits reset each month — they don’t roll over. Credit packs you buy separately never expire.

App password via IMAP. We never ask for your real password. Step-by-step →

Generate a pairing code in the dashboard, send it to our bot on Telegram, done. 30 seconds. Step-by-step →

Unlimited on every tier, with no per-inbox fee. Connect your personal Gmail, your work Outlook, your side-hustle Zoho, and your spouse’s iCloud — your agent treats them as separate inboxes and can route messages intelligently between them.

Can I use multiple social media accounts on the same platform?

Section titled “Can I use multiple social media accounts on the same platform?”

Yes. You can connect a personal X account and a business X account, for example. Your agent lets you specify which to post from.

Do I need a Discord server to use Discord integration?

Section titled “Do I need a Discord server to use Discord integration?”

No. You can DM the bot directly without having a server at all.

Tell your agent in plain English: “Every weekday at 7am, send me a morning briefing via Telegram.” Done. More →

Can my agent post to social media without my approval?

Section titled “Can my agent post to social media without my approval?”

Yes, but it’s require approval by default because the blast radius of a bad social post is high. You can loosen this per-platform. Trust levels →

  • Trial: 10
  • Standard: 25
  • Pro: 50

These are soft caps to protect scheduling performance, not your token budget. The real ceiling on “how much your agent can do per day” is your monthly credit balance — see Understanding costs.

What social platforms do you support, and how many posts can I make?

Section titled “What social platforms do you support, and how many posts can I make?”

We support X (Twitter), LinkedIn, Facebook (Pages only), Instagram (Business accounts only), Pinterest, and TikTok. All flow through the same agent workflow — draft in chat, post from a single instruction, multi-platform variants supported.

Monthly post allowance:

TierPosts/month
TrialUses the Standard allowance during the trial
Standard150 posts/month
Pro300 posts/month

Need more for a launch or burst? Grab a social post credit pack from Settings → Billing → Social post credits300 extra posts for $3.40, never expires, stacks on top of your monthly allowance. Full details →

Not yet. Voice is on the roadmap.

Does it work in languages other than English?

Section titled “Does it work in languages other than English?”

Yes. The default model handles dozens of languages well. Your agent will respond in whatever language you write to it in. Memory files can be in any language too.

On Contabo Cloud servers in the US-Central region (St. Louis, Missouri, USA). If you need EU data residency, that’s on the roadmap but not currently available.

We don’t have a formal SLA yet. Informally, we target 99.5% and have a public status page. Historically we’ve been above that.

The underlying agent framework is open source. This specific platform (the dashboard, the auth, the billing, the integrations) is not open source and not self-hostable today.

Yes. Our infrastructure is US-based but the platform is usable globally. The main caveats:

  • Payment must be in USD (Stripe handles currency conversion)
  • If you’re in a country that restricts OpenAI / Anthropic / Google, you can still use MiniMax via our default pool or BYOK

Yes — through Telegram, Discord, Slack (all of which have mobile apps), or the dashboard’s mobile web interface.

support@myagentplatform.com — we respond within 24 hours on weekdays, 48 on weekends.

For urgent issues, join the Discord and ping in #help.

Yes. Click the Ask your agent button in the bottom-right of any docs page. If you’re signed in, your own agent answers using the page you’re reading as context.

Yes — our Discord server has all our users, plus the team. Good place for questions, feature requests, and seeing what other people are building with their agents.