Bringing trust back to testimonials.
Fake reviews are everywhere. Verifiable ones are rare. We're building the tool that makes 'real' provable — without exposing the conversation.
Anyone can write a fake testimonial. Almost no one can prove a real one.
That's the entire reason Truesaid exists.
Same words. One is provable. The other isn't.
The same client could send the same thank-you. Posting it as a screenshot or quote on your website carries no weight today — anyone could have made it up. Posting it through Truesaid carries a public, scannable proof.
The "trust me" testimonial
"This service is amazing! Highly recommended to anyone."
- No way to confirm the source
- Could be invented in 30 seconds
- Screenshots can be edited freely
- Prospects scroll past it
The verifiable testimonial
"Llevábamos seis meses con la inspección bloqueada y en dos llamadas Juan lo desbloqueó. Le he recomendado ya a tres amigas en la misma situación."
- Source: real WhatsApp chat, integrity-checked
- Public verification page anyone can scan
- Synthesis grounded in chat — no invention
- Verbatim excerpts kept word-for-word
Service businesses run on word of mouth. Their testimonials don't.
Lawyers, doctors, accountants, coaches, agencies — most of their growth comes from someone telling someone else. And yet, the testimonials they show on their websites have become almost worthless: anyone can write a fake one in 30 seconds, screenshots are trivial to edit, and the prospects know it.
Meanwhile, the real proof — messages where the client actually said the work made a difference — sits unread in WhatsApp threads and email inboxes. Truesaid exists to bridge that gap, without exposing the conversation.
Concrete commitments, not slogans.
Things you can hold us to. Each number below maps to a specific piece of code or product policy.
Every Truesaid testimonial walks the same path.
From a real client conversation to a public verification page, with no human typing or invention in between.
The five rules we won't break.
Even when users ask. Even when it would be easier to ship without them. Especially then.
The text comes from the chat
The polished testimonial is a synthesis Truesaid writes from the conversation — but only of what the client actually said. Verbatim quotes are word-for-word literal. Nothing is invented, no claims are attributed that the client didn't make, no unrelated lines are stitched together.
Verification is public
Anyone can scan the QR or click the verification link and confirm the testimonial came from a real conversation. No login, no "request access", no friction. Hiding the verification would defeat the entire purpose.
The conversation stays private
Proving authenticity doesn't require exposing the chat. The verification page only shows non-sensitive metadata: import date, language, message count, integrity score. The original messages are never published. Privacy and proof are not in tension.
You can't hand-edit the testimonial
You can pick versions, change tone, regenerate, choose which excerpts to highlight — but you can't type the words yourself. If you could, the chain "real chat → faithful synthesis → public proof" would break: anyone could write whatever they wanted and slap a verification badge on it.
Your data, your control
Delete a client and everything goes: the imported chat, the testimonial, the excerpts, the public verification page, the QR. The verification URL stops resolving. No "soft delete", no archive, no hostage data.
Two outputs. Verified differently. Both real.
People sometimes assume "verified" means "every word is a literal quote". For Truesaid that's only half true — and we'd rather you know exactly what's what.
The polished testimonial
- What it is
- A short, medium, or long version that Truesaid writes from the conversation, in your chosen tone.
- What's verified
- That nothing was invented. The synthesis is grounded in what the client actually said in the chat — no claims added, no words put in their mouth.
- What it isn't
- A verbatim quote. It's a rewrite — a fair, faithful one, but a rewrite.
The verbatim quotes
- What it is
- The strongest exact phrases your client wrote, kept word-for-word with their original timestamps.
- What's verified
- 100% literal. No edits. No translation. No "polish". Pulled directly from the integrity-checked source file.
- What it isn't
- A summary or a paraphrase. If a comma is wrong, it's our bug.
The hard questions we already worked through.
The choices below shaped Truesaid. Each one is an answer we'd defend in front of a skeptical user — and several already came up in early conversations.
? Why can't I edit the testimonial myself? It would save time.
Because the moment you can edit it, the verification claim collapses: anyone could type whatever they wanted and put a "verified" badge on it. If we let users hand-edit, the chain "real chat → faithful synthesis" breaks — and a fake testimonial dressed up with a Truesaid badge is worse than no testimonial at all. You can pick versions and tone, regenerate, or choose excerpts to highlight. But the words themselves stay grounded in the chat.
? Is the verification page really public? Couldn't that leak something?
Yes, it's public. If verification weren't public, it wouldn't actually verify anything — it would just be a private claim. The page only shows non-sensitive metadata: import date, language, message count, integrity score. The chat content, the client's phone number, and any private detail are never on it. You can also hide or initialize the client's name per testimonial.
? Why don't you ask the client to approve the testimonial before publishing?
Because the privacy controls already cover that case. You decide whether to show the client's full name, just initials, or hide it entirely (blurred in exports and widgets) — per client. Adding an approval flow on top would mean storing the client's contact details, sending them links, and creating one more reason for things to stall. The "hidden / initials / full name" toggle is the consent mechanism, and it stays under your control.
? Why use AI for this?
Three reasons: quality of long-context reading (a chat can be hundreds of messages), strict instruction-following (our prompts forbid invention and the model respects that reliably), and data handling (the underlying API does not use customer data to train models). The prompt itself is conservative: if the chat doesn't justify a claim, the testimonial leaves that section empty rather than padding it.
What "Verified by Truesaid" actually proves.
Spelling out exactly what the badge means — and what it doesn't claim — keeps trust calibrated. We'd rather under-promise than overstate.
What the badge proves
- The source file passed Truesaid's integrity checks (CRC, structure, signatures).
- The polished testimonial was generated by Truesaid, with prompts that forbid invention.
- The verbatim quotes are literal — pulled directly from the source file.
- The verification page exists at a permanent public URL.
- The user who imported the chat is the one publishing the testimonial.
What the badge doesn't claim
- That every word in the polished testimonial is a verbatim quote (it's a synthesis).
- That the client gave explicit consent to publish (you control name display per client).
- That the work or claims described in the chat are objectively true — only that the client expressed them.
- That the conversation is the only one between you and the client.
- That Truesaid judged the quality of the service. We're a verifier, not a reviewer.
What Truesaid will never do.
Open standards, modern infrastructure.
Truesaid runs on Cloudflare Workers (edge compute), uses Cloudflare D1 for storage, and calls a frontier language model for the synthesis work. Every public page is rendered server-side as static HTML — indexable by Google and readable by AI crawlers.
No tracking pixels, no third-party scripts, no dark patterns. Authentication is passwordless: magic links sent to your email, valid for 15 minutes. There's no password to leak or reset.
If you want the technical specifics, the security page spells everything out — what we keep private, where data lives, what the integrity checks actually do.
Transparent answers to obvious questions.
The basic stuff people deserve to know without digging.
- Status
- Public research preview. The product is fully usable and free during beta. See pricing.
- How it makes money
- It doesn't, yet. Paid plans for high-volume use will arrive after beta — current free-tier features stay free for individual professionals.
- Where data lives
- Cloudflare's edge network (Workers + D1). No third-party trackers, no data sold or shared. Details on /security.
- How to get in touch
- Sign in and use the in-app contact, or reply to your magic-link email. Real humans on the other side.
Ready to turn a real conversation into a verified testimonial?
Sign in with just your email. No card, no password.
Get started →