Write, rewrite, and set a voice
- The method, writing task contracts, and language checks
- Everything in Core
- Same as Pro
Core is free for everyone. Pro and Studio add the full set of playbooks and market files, with more agents and tools. Prices are in US dollars. International checkout is not open yet, so for now Pro and Studio have a waitlist.
For recurring client or brand work, Pro adds a campaign command, version comparison, and batch checks. Core supports general writing and rewriting in your existing assistant.
For writing in your own voice
Free for everyone
No end date. Text under CC BY 4.0, scripts under the MIT License.
License: CC BY 4.0 and MIT
For campaigns, version comparison, and batch checks
$29 one-time
For one person, with 12 months of new versions. Each version you receive keeps working after that.
For a shared team workspace
$199 with 12 months of updates
For up to 5 people in one organization, with 12 months of new versions. Received versions remain usable for the same team. Renewal is optional.
Until international checkout opens, the Pro and Studio buttons lead to the waitlist. The opening will be announced in the public repository.
Studio adds no tools beyond Pro. It licenses the same files for up to five people in one organization.
Your AI service is separate. A Whalory purchase includes no AI subscription or credits and does not activate online generation on this website.
Pro update renewal
$15 for another 12 months of new versions, on the same Pro license key.
These prices are for buyers outside Iran. Until international checkout opens, nothing is charged: no payment, no pre-order, no deposit.
Core is free and openly licensed. The files Pro and Studio add are under a paid license that does not allow redistribution; the Core files inside them keep their open license.
Every edition includes the core method and writing task registry. Specialist references, additional tools, and ready-made install packs make the difference.
| Capability | Whalory Core | Whalory Pro | Whalory Studio |
|---|---|---|---|
| Method | |||
| Reads the assignment first Before writing, Whalory works out the task, channel, reader, and output language from the request and what surrounds it. | Included | Included | Included |
| Asks only what changes the text At most three short questions, each with a recommended option. Reply “write” and it drafts right away, with brackets for what is missing. | Included | Included | Included |
| A playbook for each task All editions include the full writing task registry and family contracts. Pro and Studio add specialist references and coordinated multichannel playbooks. Core includes the same task registry; the specialist reference library is smaller. | Limited | Included | Included |
| Campaigns and copy audits The campaign command plans multichannel work; audit finds drift in register, spelling, and variant across a folder of copy. | Not included | Included | Included |
| Facts and claims | |||
| Unknown facts stay visible Missing details go in brackets, such as [confirm: weight], and the delivery note lists what is still open. | Included | Included | Included |
| No claim without evidence An objective claim needs evidence the brand can show. Otherwise it gets a bracket, and in high-risk industries the note says “confirm before publishing”. | Included | Included | Included |
| Removes stock machine patterns Catalogs of machine patterns in English and Persian: straw-man contrasts, stacked adjectives, moral wrap-ups, dashes in the middle of a sentence. | Included | Included | Included |
| Regulation and industry notes A regulation matrix with review dates, and detailed notes per industry. Dated guidance. It is not legal advice. | Not included | Included | Included |
| Persian and English | |||
| Persian written with care Half-spaces, Persian digits, Persian quotation marks, and the Persian forms of yeh and kaf are checked by the Persian linter. Core ships the general Persian craft references; Pro adds specialist references. A checked orthography, not a claim of complete language mastery. | Included | Included | Included |
| English in the right variant US English by default. A profile or market can set en-GB, en-AU, en-CA, en-NZ, en-IE, or en-IN, with the matching spelling. | Included | Included | Included |
| Transcreation between Persian and English Copy moves between the languages as a side-by-side table with notes on what changed and why. Numbers, names, prices, dates, and conditions stay locked. | Not included | Included | Included |
| Brand voice | |||
| A voice profile for each brand VOICE.md and voice.json hold a brand’s voice and are read on every task. Core has templates, Whalory’s own voice, and two starter profiles per language; Pro has every starter. A profile is a set of instructions that travels with Whalory. It does not train a model, and a chat without files cannot remember it. | Included | Included | Included |
| Builds the profile from answers The voice workflow builds or updates VOICE.md and voice.json from three questions or from sample texts. | Included | Included | Included |
| Keeps lasting feedback Accepted feedback goes to LEARNINGS.md next to the profile and is read on every task. Chat memory is not relied on. | Included | Included | Included |
| Checks | |||
| Local checks in two languages lint.py, lint_fa.py and lint_en.py check copy on your own machine with Python 3.8 or later. They make no network calls. | Included | Included | Included |
| Says when a check did not run Without Python, Whalory checks against a manual list and writes “QA: manual” in its note. | Included | Included | Included |
| SMS length by encoding textcount.py counts SMS segments: GSM-7 allows 160 characters, or 153 per part; UCS-2, which Persian and emoji need, allows 70, or 67 per part. Providers may differ. Iranian SMS panel limits with sources are in the Pro channel data. | Included | Included | Included |
| Tools over MCP A local MCP server. Core: lint_text, lint_file, check_final, get_playbook, get_reference_section, profile_lookup. Pro adds detect_context, channel_limits, compare_texts, and ab_test_size. Six tools in Core, ten in Pro. Needs Python 3.8 or later. Local statistics are saved only after terminal opt-in. | Included | Included | Included |
| Flags changed details between versions compare_texts flags added or dropped numbers, names, quotes, and known conditions. It does not verify every fact or preserve meaning automatically. | Not included | Included | Included |
| A/B test sizing and corpus audits ab_calc.py sizes a test or checks its significance; audit_corpus.py scores a folder of copy. | Not included | Included | Included |
| Checks after each edit, if you opt in A separate Claude Code plugin lints content and locale files after each edit and never blocks one. Other hosts need a hook set up by hand. Included in the downloadable Core and Pro plugins for Claude Code. The Cursor hook has not been run inside Cursor. | Limited | Limited | Limited |
| Runs in continuous integration lint.py exits with 0, 1, or 2, so any CI system can check copy and locale files on each pull request. | Included | Included | Included |
| Review | |||
| A second reader for every draft In the plugin packages, a separate editor subagent reviews each draft blind. It sees the draft, the facts, the profile, and the format, never the writer’s reasoning. Elsewhere the same model reviews in a separate pass. | Included | Included | Included |
| Four more specialist agents Pro adds a researcher that checks facts and sources, a strategist for campaigns, a transcreator, and a voice keeper. | Not included | Included | Included |
| Channels | |||
| Channel limits with sources Character, byte, and item limits for each platform field, each with a source and a check date. | Not included | Included | Included |
| Iranian channels and marketplaces Telegram, Eitaa, Bale, Rubika, Digikala, Basalam, Divar, Iranian bulk SMS services, and ad networks. | Not included | Included | Included |
| Install packages | |||
| Prompts for any chat Ready prompts for custom instructions: 1,500 and 4,000 characters in Core, 8,000 as well in Pro. They carry a summary of the method, with no tools, memory, agents, or updates. The Core prompt pack is available on the download page. | Included | Included | Included |
| Packs for GPTs, Gems, and editors Custom GPT and Gem packs in English and Persian, a Gemini app upload, and IDE rules. The Cursor plugin is also available in Core. OpenAI now limits new GPT creation to Business, Enterprise, and Edu workspaces and plans to retire custom GPTs. | Not included | Included | Included |
| Updates and Hub | |||
| New versions Pro and Studio buyers can download new versions from the account page for 12 months with their license key. Core updates are available from the public repository. Update files manually. Hub sync and statistics sharing are disabled in this release. Local tools work without them. | Planned | Included | Included |
| Hub: rule updates and opt-in statistics Update files manually. Hub sync and statistics sharing are disabled in this release. Local tools work without them. Production trust is not approved. Built-in local rules continue to work; sharing cannot open before the trust and notice requirements are met. | Limited | Limited | Limited |
| Your files and your text | |||
| The tools stay on your machine The linters, the MCP server, and the autolint hook make no network calls. They write a file only when you run --fix, or keep weekly counts on your computer after you turn on Hub sharing. The assistant you use handles your text under its own terms. | Included | Included | Included |
| The text you write is yours In every edition the text is yours: commercial use is allowed, credit is not required, and Whalory adds no promotion of itself. | Included | Included | Included |
Reads the assignment first
Before writing, Whalory works out the task, channel, reader, and output language from the request and what surrounds it.
Asks only what changes the text
At most three short questions, each with a recommended option. Reply “write” and it drafts right away, with brackets for what is missing.
A playbook for each task
All editions include the full writing task registry and family contracts. Pro and Studio add specialist references and coordinated multichannel playbooks.
Core includes the same task registry; the specialist reference library is smaller.
Campaigns and copy audits
The campaign command plans multichannel work; audit finds drift in register, spelling, and variant across a folder of copy.
Unknown facts stay visible
Missing details go in brackets, such as [confirm: weight], and the delivery note lists what is still open.
No claim without evidence
An objective claim needs evidence the brand can show. Otherwise it gets a bracket, and in high-risk industries the note says “confirm before publishing”.
Removes stock machine patterns
Catalogs of machine patterns in English and Persian: straw-man contrasts, stacked adjectives, moral wrap-ups, dashes in the middle of a sentence.
Regulation and industry notes
A regulation matrix with review dates, and detailed notes per industry.
Dated guidance. It is not legal advice.
Persian written with care
Half-spaces, Persian digits, Persian quotation marks, and the Persian forms of yeh and kaf are checked by the Persian linter. Core ships the general Persian craft references; Pro adds specialist references.
A checked orthography, not a claim of complete language mastery.
English in the right variant
US English by default. A profile or market can set en-GB, en-AU, en-CA, en-NZ, en-IE, or en-IN, with the matching spelling.
Transcreation between Persian and English
Copy moves between the languages as a side-by-side table with notes on what changed and why. Numbers, names, prices, dates, and conditions stay locked.
A voice profile for each brand
VOICE.md and voice.json hold a brand’s voice and are read on every task. Core has templates, Whalory’s own voice, and two starter profiles per language; Pro has every starter.
A profile is a set of instructions that travels with Whalory. It does not train a model, and a chat without files cannot remember it.
Builds the profile from answers
The voice workflow builds or updates VOICE.md and voice.json from three questions or from sample texts.
Keeps lasting feedback
Accepted feedback goes to LEARNINGS.md next to the profile and is read on every task. Chat memory is not relied on.
Local checks in two languages
lint.py, lint_fa.py and lint_en.py check copy on your own machine with Python 3.8 or later. They make no network calls.
Says when a check did not run
Without Python, Whalory checks against a manual list and writes “QA: manual” in its note.
SMS length by encoding
textcount.py counts SMS segments: GSM-7 allows 160 characters, or 153 per part; UCS-2, which Persian and emoji need, allows 70, or 67 per part. Providers may differ.
Iranian SMS panel limits with sources are in the Pro channel data.
Tools over MCP
A local MCP server. Core: lint_text, lint_file, check_final, get_playbook, get_reference_section, profile_lookup. Pro adds detect_context, channel_limits, compare_texts, and ab_test_size.
Six tools in Core, ten in Pro. Needs Python 3.8 or later. Local statistics are saved only after terminal opt-in.
Flags changed details between versions
compare_texts flags added or dropped numbers, names, quotes, and known conditions. It does not verify every fact or preserve meaning automatically.
A/B test sizing and corpus audits
ab_calc.py sizes a test or checks its significance; audit_corpus.py scores a folder of copy.
Checks after each edit, if you opt in
A separate Claude Code plugin lints content and locale files after each edit and never blocks one. Other hosts need a hook set up by hand.
Included in the downloadable Core and Pro plugins for Claude Code. The Cursor hook has not been run inside Cursor.
Runs in continuous integration
lint.py exits with 0, 1, or 2, so any CI system can check copy and locale files on each pull request.
A second reader for every draft
In the plugin packages, a separate editor subagent reviews each draft blind. It sees the draft, the facts, the profile, and the format, never the writer’s reasoning. Elsewhere the same model reviews in a separate pass.
Four more specialist agents
Pro adds a researcher that checks facts and sources, a strategist for campaigns, a transcreator, and a voice keeper.
Channel limits with sources
Character, byte, and item limits for each platform field, each with a source and a check date.
Iranian channels and marketplaces
Telegram, Eitaa, Bale, Rubika, Digikala, Basalam, Divar, Iranian bulk SMS services, and ad networks.
Prompts for any chat
Ready prompts for custom instructions: 1,500 and 4,000 characters in Core, 8,000 as well in Pro. They carry a summary of the method, with no tools, memory, agents, or updates.
The Core prompt pack is available on the download page.
Packs for GPTs, Gems, and editors
Custom GPT and Gem packs in English and Persian, a Gemini app upload, and IDE rules. The Cursor plugin is also available in Core.
OpenAI now limits new GPT creation to Business, Enterprise, and Edu workspaces and plans to retire custom GPTs.
New versions
Pro and Studio buyers can download new versions from the account page for 12 months with their license key. Core updates are available from the public repository.
Update files manually. Hub sync and statistics sharing are disabled in this release. Local tools work without them.
Hub: rule updates and opt-in statistics
Update files manually. Hub sync and statistics sharing are disabled in this release. Local tools work without them.
Production trust is not approved. Built-in local rules continue to work; sharing cannot open before the trust and notice requirements are met.
The tools stay on your machine
The linters, the MCP server, and the autolint hook make no network calls. They write a file only when you run --fix, or keep weekly counts on your computer after you turn on Hub sharing. The assistant you use handles your text under its own terms.
The text you write is yours
In every edition the text is yours: commercial use is allowed, credit is not required, and Whalory adds no promotion of itself.
Give the product features, price, audience, voice, and intended channels.
The campaign command guides your assistant through planning, drafting, and reviewing website copy, a social caption, and a text message.
The comparison tool flags changes to numbers and names. You still check the meaning and verify the facts.
This is an illustrative workflow. Your assistant generates the copy; quality depends on your information and the model.
Prepare your own campaign briefAfter payment confirmation, the result page shows your key and files. Your license page can issue fresh download links.
Follow its install guide, or supply the instructions and knowledge files. Tool execution depends on the host.
The purchase includes twelve months of updates. Versions you receive remain usable afterward. Renewal is optional.
Local tools require Python. GPT and Gem packs depend on account eligibility. OpenAI currently excludes personal accounts from creating new GPTs; check your assistant’s compatibility before buying. Official OpenAI guidance · Compatibility and install guides
The website workspace is separate from the paid Studio license. It supports briefs, voice profiles, editing, and exports. Online generation is currently disabled.
Pro is a one-time purchase for one person. Studio covers up to five people for a year of updates. In both editions, versions you receive remain usable after that period. Renewal is optional; there is no automatic charge.
Whalory sells no access to any assistant. Whatever you pay for the AI you use stays separate, under that service’s terms.
You can request a full refund within seven calendar days of a purchase, even after downloading the files. Submit a private support request. The refund is processed within seven working days after the buyer and order are confirmed.
It has no opening date yet. Core is free everywhere, today.
A separate hosted service. Your downloadable Core, Pro, and Studio licenses keep their existing terms.
$9 USD per month
$19 USD per month
$49 USD per month
Longer text, higher quality, and multiple outputs use more credit. You review the maximum before each generation. No unlimited AI plan. Renewal is manual.
Online checkout is not available yet. Purchases open only after the payment provider and service settings are verified.
See the online workspaceCore gives you Whalory’s writing method inside your own AI tool; it does not include hosted AI processing. Online plans bring server-side generation, shared brand information, saved documents, team access, and a defined credit allowance into one account. You do not need to enter your own API key. Choose it when keeping the work together saves you time.