# How Much an API-Generated Presentation Costs

> render.ai charges 40 credits for a presentation generated over its MCP server or REST API, and generation is the only call either interface meters.

Canonical: https://render.ai/guides/api-generated-presentation-cost
Last updated: 2026-09-27

render.ai charges one flat rate for a presentation generated over its MCP server or REST API, and
generation is the only thing either interface meters. That rate does not move with the length of the
deck, the number of charts on it, or the images inside it. Everything an agent does around a
generation - listing themes, watching the run, reading the balance, calling it off - is free.

Whether it is a weekly metrics deck, a client proposal, an investor update, or a one-off board pack,
every deck occupies the same line in the budget. A predictable per-deck figure is what a developer
needs before an agent that produces slides goes near production, and what an operations lead needs
before a reporting job is scheduled to run every Monday. Pricing an agent workflow on render.ai
means counting presentations, not calls, tokens, slides, or seconds.

**Key facts**

- render.ai charges 40 credits per presentation created, at the same price over the MCP server and
  the REST API.
- Every call around a generation is free on render.ai: the theme list, each status poll, the usage
  check, and a cancel.
- Hands-on edits to a generated render.ai deck cost 0 credits, so an agent's output can be corrected without
  a second charge.
- The render.ai free plan grants 400 credits once at signup, which covers 10 presentations from that
  grant.
- An agent can spend credits over render.ai's MCP server on any plan, and over the REST API on a
  paid plan.

## render.ai Charges One Flat Rate Per Presentation

render.ai charges 40 credits per presentation created, and the figure is identical over the MCP
server and over the REST API. A 10-slide deck and a 40-slide deck both cost 40 credits, because the
charge is per presentation rather than per slide. There is no volume discount, no cheaper draft
mode, and no per-call fee layered on top.

Credits are the only unit of cost in the render.ai API. The number of presentations an agent can
still create is its credit balance divided by 40. render.ai deducts the 40 credits once a presentation completes fully. A
generation ending at `failed`, or one an agent stops with `cancel_generation`, costs nothing.

## One Charge Covers Everything render.ai Reads Out of the Prompt

render.ai reads the whole of `inputText` and works out for itself which points become slides, which
become bullets underneath them, and which figures are worth a chart - all inside the single
40-credit charge. The agent sends prose and gets back a deck's worth of structure. None of that work
is billed separately, and there is no cheaper mode that does less of it.

Charts are part of the same 40 credits. An agent that sends monthly revenue, a customer split, or a
ranking gets back a line, donut, or bar chart drawn in the deck's theme, with no charge per chart.
A deck with six charts costs exactly what a deck with none does.

`textMode` decides how much of the source survives, and none of its values changes the price.
`condense` shortens supplied text, `preserve` keeps it as written, and `generate` writes slides from
a topic. All three produce a presentation at 40 credits.

## One Charge Returns a Finished, Themed Deck

render.ai returns a complete presentation for that one charge: headline, supporting points, layout,
imagery, and charts, with a theme applied across every slide. Professional visual quality is the
default rather than a pass somebody makes afterwards, and the layout is generated for that deck
rather than poured into a fixed template. `list_themes` enumerates the themes an agent can pass as
`themeId`, and calling it costs nothing.

Hands-on edits in the render.ai editor afterwards cost 0 credits. Rewriting slide text, deleting
slides, reordering them, changing fonts and colours, and swapping the theme deck-wide all happen
without a second charge. A near-miss is therefore cheap to fix, and a regeneration is not.

Every render.ai export is free of credits, in every format. PDF, a published web link, PowerPoint (`.pptx`),
and Google Slides are all free of credit charges once the deck exists; which of them an account can
reach depends on the plan rather than on the balance.

## Budgeting an Agent's render.ai Credits: Complete Guide

Budgeting an agent's render.ai credits takes 8 steps, from a cold start to a deck delivered inside
the budget. Steps 1 to 3 all happen before anything is spent.

1. Call `check_usage` before generating, and treat the balance it returns as a gate rather than a
   log line.
2. Divide that balance by 40 credits per presentation to get the number of generations still
   available.
3. Stop and report "not enough credits for this deck" when the balance is below 40, instead of
   letting the generation call fail partway through a longer task.
4. Write `inputText` once and carefully, because it is the only content the render.ai generator
   receives and the only lever on quality that costs nothing.
5. Set `numCards`, `themeId`, and `textMode` on that same call, since none of them changes the
   40-credit charge.
6. Call `generate_presentation` and keep `generationId` and `contentUrl` from the response.
7. Edit the deck by hand in the render.ai editor, at 0 credits, when the first result is close enough, rather
   than generating a second one.
8. Call `cancel_generation` on a run nobody wants any more, which stops it before it completes and
   is therefore not charged.

Speculative regeneration is the main way an agent wastes an account. Generating the same deck 4
times creates 4 presentations and spends 160 credits, which is 40% of the free plan's 400-credit
signup grant. An edit in the render.ai editor changes exactly the slide that is wrong and costs
nothing, so a second generation earns its keep only when the input itself was wrong.

## Only Generation Is Metered on the render.ai API

render.ai meters exactly one of the 5 tools on its MCP server, and that tool is
`generate_presentation`. The other 4 are free however often an agent calls them.

| Tool | Cost |
| --- | --- |
| `generate_presentation` | 40 credits per presentation created |
| `list_themes` | 0 credits |
| `get_generation_status` | 0 credits |
| `check_usage` | 0 credits |
| `cancel_generation` | 0 credits |

Because those 4 calls are free, an agent can lean on them:

- **Poll freely.** `get_generation_status` can run until the generation reaches a terminal state,
  at render.ai's documented cadence of once every 10 seconds.
- **Check first.** `check_usage` reads the balance before generating, so the agent knows its budget
  up front.
- **Resolve themes live.** `list_themes` returns the current `themeId` values on every run.
- **Cancel early.** `cancel_generation` stops a run before it completes. A run that never completes
  is never charged.

## How Many Presentations Each render.ai Plan Covers

render.ai plans cover 10, 25, and 125 presentations. Each figure is the plan's credit allowance
divided by 40 credits per presentation created.

| Plan | Price (USD per month) | Credits | Presentations (at 40 credits each) | Slides per prompt (maximum) |
| --- | --- | --- | --- | --- |
| Free | $0 | 400 at signup | 10 | 10 |
| Pro | $28 | 1,000 per month | 25 | 20 |
| Business | $56 | 5,000 per month | 125 | 60 |

The free plan's 400 credits are granted once at signup rather than refreshed monthly, so those 10
presentations are a starting grant rather than a monthly allowance. Free accounts can earn further
credits through promotions such as inviting members. An agent on a free account should therefore
read the live balance with `check_usage` instead of assuming a fixed lifetime budget. Pro and
Business credits refresh every month.

Running out is a clean failure rather than a degraded deck. The render.ai generation call returns an
out-of-credits error carrying an upgrade action, and the right response is to present that link and
stop calling the tool rather than retrying. Rate limits are enforced separately and return
`RATE_LIMIT_EXCEEDED`, but credits are normally what a workflow meets first.

| Plan | Generations per hour | Concurrent generations |
| --- | --- | --- |
| Free | 30 | 2 |
| Pro | 100 | 5 |
| Business | 500 | 15 |

## render.ai's MCP Server Works on Every Plan

render.ai's MCP server lets an agent generate decks from the first minute of a brand new account,
on every plan. The REST API serves teams on paid plans who call render.ai from their own code.

| Interface | Endpoint | Plans |
| --- | --- | --- |
| MCP server | `https://api.render.ai/mcp` | Every plan, free included |
| REST API | `https://api.render.ai/v1` | Paid plans |

MCP is the natural starting point for most agent workflows. Adding the connector is itself the
sign-up path, so an agent can bring someone new to render.ai straight to a first deck. The
credentials are issued during authorization, so nobody creates, copies, or pastes an API key.

That first connection takes one browser sign-in and one consent click from a person. Every call
after that belongs to the agent.

## What a Paid render.ai Plan Adds to the Output

A paid render.ai plan delivers unbranded output, on the API and in the editor alike, and adds
PowerPoint and Google Slides export in the editor.

- **PDF** - An agent requests it with `exportAs: "pdf"`, and a person can export it from the
  editor. Paid plans deliver it without branding.
- **Live website** - A person publishes it from the editor on any plan, unbranded on paid plans.
- **PowerPoint (`.pptx`) and Google Slides** - A person exports them from the editor on a paid plan.

Every export is free of credits, so the plan decides the branding and the formats while the balance
stays reserved for generation. An agent asked for a `.pptx` hands over the `contentUrl` from the
generation, and the person exports the file from the editor.

## Where to go next

The full agent-facing documentation is at [render.ai for agents](/agents.md), with costs and limits
at [pricing for agents](/pricing.md). Human-readable plan details are on [pricing](/pricing), the
product overview is at [render.ai AI presentations](/lp/presentations-ai), and related articles are
collected under [agents and API](/guides/agents).