How Story Matching Works
When you share a story, Our Shared Stories quietly looks for other stories that feel related — not by exact words, but by meaning, themes, and tone. This is done using a modern language technique called embeddings.
What Are Embeddings?
An embedding is a kind of “mathematical fingerprint” of a story. Instead of simply counting words, the system converts your story into a numeric pattern that captures its overall meaning — for example whether it speaks about loss, separation, imprisonment, displacement, injury, or recovery.
These fingerprints are not readable as text. They cannot be used to reconstruct your story, and they do not contain your name, email, or any other personal details.
How We Find Similar Stories
When you publish a story, we generate an embedding and store it securely. When you open your own published story, the system compares its embedding with those from other stories and measures how “close” they are in meaning.
We then show you up to a small number of stories that are closest in meaning to your own. These might share:
- Similar types of events (for example, detention, injury, or loss)
- Similar emotional tone (for example, grief, uncertainty, or resilience)
- Similar journeys (for example, living with loss, rebuilding, or returning home)
This is not a “perfect match” or a diagnosis. It is simply a gentle way to help you find stories that may resonate with what you have lived through.
How Different Languages Are Handled
People write in many languages: Arabic, English, and others. Our Shared Stories uses multilingual embeddings, which are designed to understand meaning across languages.
In practice, this means:
- Stories written in different languages can still be recognised as similar in meaning.
- The system first prefers stories written in the same language as yours, where possible.
- Where it is helpful, it can also draw on stories written in other languages that express similar themes.
Your story is always shown exactly as you wrote it. We do not automatically translate your text.
Which Stories Are Included
Embeddings are generated only for stories that you choose to publish. Drafts and private stories are not used for story matching.
You can only see similar stories for your own published stories. Other people cannot request “similar stories” based on your writing.
Privacy and Safety
Our Shared Stories is built for emotional safety first. Embeddings are one way we can gently connect related stories without turning the platform into social media.
- We only generate embeddings from the story text you choose to publish.
- We never include your name, email, or account details in matching.
- Embeddings cannot be used to reconstruct your original story.
- All data is transmitted over encrypted connections and stored securely.
If at any point you unpublish or delete a story, it will no longer be used in future similarity matching.
A Tool to Support, Not Judge
Story matching is not an assessment of you or your experience. It is simply a quiet tool that helps you see that others have faced related struggles, and that your story sits alongside many others in a shared space of care and respect.