Pricing
Pay for the workspace. Bring your own model.
ModMail does not resell model access: you connect your own provider key and they bill you directly, at their rates, with no markup. What the plan pays for is the workspace — mailboxes, agents that work your inbox, and the assistant.
14-day free trial · one agent · no card
Free
One inbox, your own AI key.
- 2 mailboxes
- AI assistant with your own provider key
- All hosted providers: OpenAI, Anthropic, Google, OpenRouter
- 50 emails a day
- Agent Mode and local endpoints are not included
Starter
For one person with a few addresses.
Save $2/monthlaunch price
- 3 mailboxes
- 1 agent — sales or marketing
- AI assistant, 100 emails a day
- All providers, hosted and local
Pro
Most chosenFor someone who runs more than one conversation at once.
Save $7/monthlaunch price
- 5 mailboxes
- 2 agents — e.g. sales and marketing
- AI assistant, 500 emails a day
- All providers, hosted and local
- 3% off when paid yearly
Max
The whole roster.
Save $20/monthlaunch price
- 10 mailboxes
- Every agent role, including ones we ship later
- AI assistant, 1,000 emails a day
- Every agent role, including new ones
- 10% off when paid yearly
Enterprise
Everything in Max, with room for more.
- Everything in Max
- Unlimited mailboxes and agents
- Higher sending and agent usage
- Onboarding and a direct line to the team
| Usage | Free | Starter | Pro | Max | Enterprise |
|---|---|---|---|---|---|
| Mailboxes | 2 | 3 | 5 | 10 | Unlimited |
| Agents | — | 1 | 2 | Unlimited | Unlimited |
| Emails a day | 50 | 100 | 500 | 1000 | Unlimited |
| Hosted AI providers | Yes | Yes | Yes | Yes | Yes |
| Local / custom endpoints | — | Yes | Yes | Yes | Yes |
What it costs to run
A day of heavy personal use — summaries, drafts, a few agent runs — lands in the low cents on a small hosted model and can exceed a dollar on a frontier model. Triaging with a cheap model and drafting with an expensive one is the usual arrangement, and it is why ModMail keeps providers per user rather than picking one for you.
If inference must not leave your network, a local model costs nothing per token beyond the machine it runs on. Local and custom endpoints (Ollama, LM Studio, vLLM) are included from Starter upward; Free covers the hosted providers.
Start with the trial
Two weeks with the assistant, every provider and one agent running on your mailbox — no card. If it earns its place, Starter is $8 a month; if it does not, your mail is still yours and nothing is charged.