Skip to content

Plans explained

Three tiers. Prices are current as of this writing; check myagentplatform.com for the latest.

  • Who it’s for: Everyone starts here
  • What you get: Full access to Standard tier features for 15 days
  • What it costs: A $10 deposit that’s credited back toward your first real month if you subscribe. If you don’t subscribe, you keep the $10 of usage credits you consumed during the trial.

The trial is genuinely unlimited in terms of capabilities — you’re not seeing a “lite” version. It’s the full Standard tier with a time limit.

Standard — $29/month (or $24/mo if billed annually)

Section titled “Standard — $29/month (or $24/mo if billed annually)”

The default choice for most users.

Standard
Price$29/mo or $290/yr ($24/mo equivalent — 2 months free)
Included AI usage~5M MiniMax M2.7 tokens/month (enough for typical daily use — see below)
ModelMiniMax M2.7 (or bring your own key — see below)
Sub-agents5 concurrent
Storage10 GB workspace
Scheduled tasks25 cron jobs
Web search credits100 searches/month
Social posts150 posts/month
Messaging channelsTelegram, Discord, Slack
Email & CalendarIncluded — unlimited inboxes
SupportPriority community + email

Best for: Solopreneurs, freelancers, and small business owners using their agent as a personal assistant for daily work.

Pro — $59/month (or $49/mo if billed annually)

Section titled “Pro — $59/month (or $49/mo if billed annually)”

The power-user tier.

Pro
Price$59/mo or $590/yr ($49/mo equivalent — 2 months free)
Included AI usage~15M MiniMax M2.7 tokens/month (3× Standard)
ModelMiniMax M2.7 (or bring your own key — see below)
Sub-agents10 concurrent (2× Standard)
Storage25 GB workspace (2.5× Standard)
Scheduled tasks50 cron jobs (2× Standard)
Web search credits300 searches/month
Social posts300 posts/month
Messaging channelsTelegram, Discord, Slack
Email & CalendarIncluded — unlimited inboxes
Priority queueYour agent’s fires jump ahead of Standard-tier traffic
SupportPriority email, response within 24 hours

Best for: Heavy users who run lots of scheduled tasks, do deep research across multiple connected inboxes, or need the bigger workspace and cron limits.

Pro isn’t about unlocking features — every tier has the full agent, BYOK, channels, and the autonomy layer (Standard+). Pro is about headroom for when you’ve outgrown the limits:

  • You’re running 25+ scheduled tasks. Standard caps at 25, Pro at 50.
  • You run through your included AI usage mid-month. Pro gives you 3× as much ($15 vs $5).
  • You want a bigger workspace. Pro has 25 GB vs Standard’s 10 GB.
  • You’re spawning 5+ sub-agents per research run. Standard caps at 5 concurrent; Pro doubles that.
  • You want your fires to skip the queue. Pro gets priority scheduling during high-load periods.
  • Monthly — charged on the same date every month, cancel anytime
  • Annual — charged once per year, pay for 10 months and get 12 (≈17% cheaper), cancel anytime (refund prorated)

You can switch between monthly and annual at any time from Settings → Billing. Switching mid-cycle prorates the difference.

Here’s how AI usage actually works on this platform, without the marketing spin:

Every message your agent processes consumes some amount of LLM tokens. Tokens are the atomic unit that model providers bill for — roughly, one token is three-quarters of a word. We pay for them and pass the cost along.

Your plan includes a monthly token budget priced against MiniMax M2.7 — roughly 5 million tokens on Standard, roughly 15 million on Pro. Tokens are the honest unit because the dollar value of a token shifts as model prices change. What stays constant is the amount of work you can get done:

  • A quick chat (“summarize this article”) → a few thousand tokens
  • A full inbox triage with drafts → tens of thousands of tokens
  • A deep research task spawning sub-agents → a few hundred thousand tokens
  • A 30-minute coding session with heavy browser + tool use → a million or more

Real-world usage: most solo users on Standard finish the month with tokens to spare. Users who run a lot of deep research, code generation, or heavy scheduled tasks tend to top up or step up to Pro.

Why we talk in tokens, not dollars. We used to quote AI usage in dollar amounts (“$5 of tokens per month”). We moved away from that because (a) model prices drop over time so the same dollar buys more, and (b) it gave people the misleading impression that their included usage was currency they could spend on things other than model calls. Tokens are the thing you’re actually buying, so tokens are what we quote.

Your agent doesn’t stop working — it warns you at 75% and again at 95% of your included usage. If you actually hit zero:

  • Option 1 — Credit packs. Buy a top-up anytime from the Credit packs card in Settings → Billing. Unused credit rolls over indefinitely and stacks on top of next month’s included amount.
  • Option 2 — Bring your own key (BYOK). Available today. If you’d rather pay the model provider directly, plug in your own API key (Anthropic, OpenAI, DeepSeek, OpenRouter, MiniMax, etc.) and your agent uses that key instead of our metered pool. Set it up from Settings → Agent → Bring Your Own Key.
  • Option 3 — MiniMax coding plan. MiniMax sells a flat-rate subscription for M2.7 that’s ideal for users who run many coding or research sessions per day. Subscribe to their plan directly, drop the key into BYOK, and you effectively get unlimited MiniMax M2.7 usage for a fixed monthly fee that bypasses our included-usage budget entirely.

BYOK is not a premium feature. It’s available on every tier, including trial. Use whichever combination works — platform-metered, your own key, or both with automatic fallback.

See Upgrading and downgrading →.

See Canceling your subscription →.