Zro

A closer look at Zro

Zro by MoonMath.ai is private AI inference for coding agents: every Zro model, paid as you go, with no subscription, zero request retention, and no training on requests, through an OpenAI-compatible API and a CLI launcher. A free plan covers daily usage on Flash models, Pro and Max add monthly usage budgets for everyday and heavier agent work, and one-time usage packs top up any plan. Agent tools such as web search and fetch are billed per call outside token estimates.

Editorial guidance based on published documentation. The workflow below is a suggested starting point, not a measured product result.

When to choose Zro

• Coding-agent users who want private inference without a subscription, since Zro charges by usage packs with no recurring charges and zero request retention.
• Everyday coding-agent users who want a monthly budget, since Pro includes a $40 monthly usage budget (or $16 per month billed yearly) up to 1B tokens.

When to choose something else

• Buyers who want a subscription-only service with no pack expiry, since usage packs expire after 90 days and have no recurring charge.
• Buyers who want prorated refunds on cancellation, since the pricing page states cancellation takes effect at the end of the current billing period with no prorated refunds.

A practical starting workflow

1. Install the CLI launcher and log in, then launch a coding agent such as Claude Code through Zro.
2. Run everyday agent work on the Free daily Flash usage or on Pro's monthly budget, with models including DeepSeek V4.1 Flash, GLM-5.3, and Dolly security models.
3. Add usage packs of $5 to $100 when more budget is needed, without a subscription.
4. Use agent tools per call for live web search and URL fetching when the agent needs fresh context.

Check before committing

• Amounts are shown with a $ sign but no named currency on the pricing page, so they are recorded as observations, not as currency-named facts.
• Token totals assume DeepSeek V4.1 Flash with typical cached agent usage; other models use the budget faster and actual totals vary by prompt shape, output length, and cache reads.
• Yearly plans are billed annually and provide the same monthly spend budget as monthly plans; monthly spend resets on the same calendar cadence for both.

Common questions

Compare approaches for this task

• Zep — Enterprise teams building agents that need reliable, governed context, since Zep unifies business data, documents, and conversations into shared context with graphs, batches, and an MCP server.; Teams that want credit-based self-serve pricing with a trial, since Flex and Flex Plus list included credits, top-ups, and rollover.
• Zinc — Developers and AI agents that want to buy anything online without building retailer integrations, since Zinc covers discovery, ordering, tracking, and returns in one API.; Marketplaces and rewards programs that place orders at volume, since Zinc offers volume pricing from a thousand orders a month and a wallet funded once and spent down.