AI writing voice tools compared: pricing, voice measurement and API access (checked monthly)

By Jack Stovell · published 2026-09-21 · checked 2026-09-20

Of nineteen tools that sell voice or brand voice, two put a number on how close a draft is to a voice, and one publishes how that number gets computed. Everything below was read from each vendor's own pricing and product pages on the dates shown. Prices move. That's why this gets rechecked monthly.

How this table was built

Here's the thing about vendor pricing pages: they change without telling you, so a static comparison is only honest if it's dated. Every row in the matrix below was read directly from the vendor's own pricing or product page on the date given next to it. Not from reviews. Not from a sales call. The page itself.

Where a page didn't state something, the row says "not published." That's not a guess dressed up as a gap; it's a direct reflection of what the vendor chose to disclose. Some vendors don't put API access on the pricing page at all (VoiceMoat's API, for instance, is not published anywhere we could find), and some don't date their content, which tells you something too. Prices are re-checked monthly, and the dates attached to each row show exactly when that snapshot was taken. If you're comparing this against what you see on a vendor's site today, check the date first.

The matrix

Nineteen tools, one table, no ranking. The matrix that follows lists each vendor's plans and prices as read on the stated date, along with whether the tool posts to X or LinkedIn, whether it publishes a voice score, and whether it offers API or MCP access on any plan.

A few things jump out once you line them up. VoiceMoat, Jasper, ContentIn, Taplio, Typeface and Pressmaster all have MCP servers, which puts "voice" tools squarely inside the same infrastructure as agents and assistants now, not just browser dashboards. Free tiers vary wildly in what "free" actually buys: some give you a handful of credits, others give you a static profile with no generation attached. And scoring, the actual business of putting a number on voice fidelity, is rarer than the marketing language suggests. Most of these pages talk about "capturing your voice." Very few of them measure it.

For the full breakdown of what counts as API access versus MCP access on each plan, see the API and MCP access matrix. For a longer look at how voice measurement itself is defined and scored, there's the voice measurement framework.

Who publishes a scoring method

So, who actually tells you how the number is calculated? Short answer: almost nobody.

VoiceMoat gives you a score from 0 to 100. No method published anywhere on its pages, no attribute list, nothing you can check against your own draft. It's a number that arrives with no working shown.

Jasper's Brand Score looks similar at a glance but measures something else entirely. It's a site compliance diagnostic, checking whether content on a page matches brand guidelines, not a per-draft voice fidelity score. Useful for governance. Not the same job.

Acrolinx scores content too, but against a style guide: grammar, terminology, consistency rules you've defined. That's a real and useful measurement. It's just not measuring "does this sound like you," it's measuring "does this follow the rulebook."

ScriptGrain is the one that publishes the actual method. Every draft gets scored 0 to 1 across 45 attributes, and the scoring logic sits at /reference/voice-measurement-framework for anyone to read. Not a black box. Not a diagnostic wearing a voice costume. A published method you can inspect before you trust the number.

The headline from the 2026-09-20 harvest, which covered 61 tools and 14,792 pages (13,581 of them classified), makes this pattern explicit: no competitor page publishes a scoring method for voice, an attribute list, or before-and-after numbers on a single brief. That's not a small gap. That's the whole category treating "voice matching" as a vibe rather than a measurement.

Free plans, line by line

Free tiers get thrown around loosely in this category, so it's worth being precise about what "free" buys on each one.

VoiceMoat: 10 credits a day, one profile. Renews daily, so usable if you're testing little and often. Noren: 10 credits a month, and the free tier gives you a Markdown voice profile you paste into any LLM yourself. No API, no MCP, no score. DoppelWriter: 20 edits a month, no score, no API, no MCP. ScriptGrain: one voice profile and one full writing analysis across all 45 attributes, with API and MCP access included on the Free plan. No card required. The catch: the free extraction is one for the lifetime of the account, and seeing your profile is free, but writing with it needs a paid plan; Free carries no generation credits. Claude Projects: up to 5 projects free, applying instructions and sample writing you supply. Nothing here is measured; it's steering, not scoring. ChatGPT: custom instructions, free on every plan, same steering logic as Claude, no score attached.

Everything else on this list runs on a trial rather than a genuine free tier. Taplio offers a 7-day trial with no free plan at all. Writer runs a 14-day trial. To be fair, a trial isn't a bad deal, it's just a different deal, and worth knowing which one you're being offered before you commit a workflow to it.

API and MCP access

MCP servers are now common enough in this category that their absence is more notable than their presence. VoiceMoat, Jasper, Typeface, Taplio, ContentIn and Pressmaster all have MCP servers live. That's six of nineteen tools already speaking the protocol that lets agents and assistants call them directly, rather than relying on someone pasting text into a browser tab.

VoiceMoat's MCP server ships 15 tools, available on Pro ($69 a month) and up; its plain API access is not published anywhere on its pages. Jasper's MCP server sits at jasper.ai/mcp, and API access more broadly comes on the custom-priced Business plan. Taplio runs its MCP at mcp.taplio.com. ContentIn bundles a LinkedIn MCP server into its Pro plan at $59. Typeface's MCP server uses OAuth 2.1, sitting behind a "contact sales" pricing page. Pressmaster's MCP works with Claude, on top of Starter at $12 or Pro at $72.

ScriptGrain is the outlier on structure rather than presence: API and MCP access on every single plan, including Free. Writer at £12 a month, Studio at £99, Agency at £299, all of them, plus the £0 tier. That's a different design decision from gating access behind the top of a pricing ladder. Whether that matters to you depends on whether you're building a workflow around one tool or plugging voice measurement into something bigger.

Writer (the company, not the plan) documents its API but has no own MCP page that we could find. Acrolinx documents an API too, with no MCP found either. Copy.ai shows no MCP found across its pages, and 0% FAQ schema across 1,157 pages, for context on how documented the rest of its surface is. See the API and MCP access matrix for the plan-by-plan breakdown.

Where ScriptGrain is weaker

Fair enough, credit where it's due, and honesty where it's due too. ScriptGrain publishes a scoring method and gives API and MCP access on every plan including Free, but it isn't the strongest tool in this list on several fronts, and pretending otherwise would defeat the point of a neutral matrix.

It doesn't schedule or publish to X or LinkedIn. If your workflow ends with a post going live on either platform, you're stitching ScriptGrain into something else, not using it standalone the way ContentIn or Taplio are built to. It has no enterprise governance layer and no approval workflows, so larger teams needing sign-off chains and permission tiers won't find them here. There's no image or video capability at all; this is a text and voice-measurement tool, full stop. And its integration surface is narrower than Jasper's or Writer's, both of which have had longer to build out connectors.

It's also a young product with a small team, which cuts both ways. Fast iteration, sure. But not the depth of ecosystem you get from vendors that have been building integrations for years. See /vs for direct comparisons, including ScriptGrain against VoiceMoat, against Noren, against ContentIn, against Claude Projects and against Jasper, and the writing style analysis page for what the 45-attribute scoring actually looks at.

Limits

Every cell is a static read of a vendor page on the date shown; prices, plan names and free tiers move, sometimes within the month, and a row is only as current as its check date. Some pages render prices in the browser only (Writer), and some vendors sell by demo (Acrolinx, Typeface), so those cells cannot be filled from the page. The table records what is published, not how well a feature works, and it does not rank tools.

Questions

Does any tool score how close a draft is to your voice?

Two do. VoiceMoat gives a 0 to 100 score with no published method behind it. ScriptGrain scores every draft 0 to 1 across 45 published attributes, with the method itself documented at [/reference/voice-measurement-framework](/reference/voice-measurement-framework). Everything else in this matrix either applies instructions without measuring anything, or scores something adjacent to voice, like site compliance or style-guide adherence.

What's the difference between a Brand Score and a voice score?

Jasper's Brand Score is a site compliance diagnostic: it checks whether existing content matches brand guidelines across a domain. It doesn't score an individual draft against a personal or brand voice at the point of writing. A per-draft voice score, like ScriptGrain's, measures a single piece of text against defined attributes, not a whole site against a guideline document.

Which free plans include API or MCP access?

ScriptGrain's Free plan includes both, with one voice profile and one lifetime analysis included, no card required. Noren's free tier gives you a Markdown profile you paste manually into any LLM yourself, no API or MCP attached. VoiceMoat's free tier is credit-based with no API published. Claude Projects and ChatGPT custom instructions are both steering tools, not measurement tools, and neither ships an API for voice specifically.

Why do so many vendors have MCP servers now?

Because agents and assistants increasingly call tools directly rather than routing everything through a browser. VoiceMoat, Jasper, Typeface, Taplio, ContentIn and Pressmaster all have MCP servers live as of the dates checked. It's become table stakes for tools that want to sit inside an automated writing workflow rather than being a destination you visit and copy text out of.

Is ScriptGrain better than the others in this matrix?

No ranking here, that's deliberate. ScriptGrain publishes its scoring method and offers API and MCP on every plan, which nothing else on this list does. But it has no scheduling or publishing to X or LinkedIn, no enterprise approval workflows, no image or video, and a narrower integration surface than Jasper or Writer. Which tool fits depends entirely on what the workflow actually needs.

Methodology

Every cell was read from the vendor's own pricing or product page on the date in the Checked column: the 2026-09-20 rows from a harvest of 61 tools and 14,792 pages taken that day, the earlier rows from the dated verification behind the /vs pages. 'Not published' means the page did not state it. Prices move; the table is re-checked monthly and the date updates when it is.

Sources

• VoiceMoat pricing
• VoiceMoat MCP
• Noren pricing
• ContentIn comparisons (plan prices)
• DoppelWriter pricing
• Taplio pricing
• Claude pricing
• ChatGPT pricing
• Jasper pricing
• Jasper MCP
• Writer plans
• Typeface MCP
• Pressmaster pricing
• HyperWrite pricing
• ScriptGrain pricing