DEV Community: Anistratenco

What a health copilot must do before it earns your trust

Anistratenco — Tue, 22 Sep 2026 12:47:32 +0000

Health Assistant is not defined by a chat box. It is defined by the jobs around the chat that make an answer useful and correctable.

The first job is context. A person can ask about sleep, training, supplements, a result or a routine, but the same words mean different things in different lives. The product needs to keep the user’s chosen context connected to the question while avoiding assumptions that were never confirmed.

The second job is correction. Health information changes, priorities change and people remember details differently. A useful system must let someone correct an earlier statement, change the importance of a goal and see that correction reflected later. Continuity without correction would simply make an old mistake more persistent.

The third job is explanation. Health science is not a single certainty switch. People need an answer that distinguishes what is established from what is uncertain, and they need that distinction in language they can use. The product’s role is to help a person understand a concern and prepare for better decisions, not to quietly present a diagnosis or replace a qualified professional.

The fourth job is handling different kinds of input responsibly. A future health workspace may include documents, voice and other user-selected material. Each input raises a different question about consent, temporary analysis, retained context and deletion. A file that helps answer one question should not silently become a permanent fact about the person.

The fifth job is recovery. Conversations can be interrupted, long tasks can take time and an answer can be incomplete. The user needs to know what was completed, what remains uncertain and what action is waiting for them. A graceful failure is part of product quality, especially when the subject is health.

The current public beta is focused on the conversational health workspace and the context a person chooses to share. More proactive work, deeper file support, voice, and mobile or watch experiences are planned phases, not claims about everything available today.

That scope is deliberately human. A health copilot should make science easier to use without making the person smaller than the system. It should help someone organise a concern, understand the terms, revisit a decision and say, “That is not right for me.” If it cannot make room for that sentence, it has not earned trust yet.

https://anistratenco.com

healthai #healthtech #digitalhealth #preventivehealth

What we mean when we talk about file support

Anistratenco — Tue, 22 Sep 2026 09:39:36 +0000

Finding the right document can take longer than asking the question. A report might be in an email, an export in a downloads folder and an older version somewhere on another device. We want the health workspace at Anistratenco to make that part of looking after yourself less awkward.

Health Assistant starts with a conversation. The file experience we are developing around it should help someone move from having a document to understanding the part they care about. That could mean discussing unfamiliar language in a report or returning to a question after receiving another document.

The planned Health space is the next part of this work. Our public site currently lists October 7 for the records and exports experience. We want it to give the material behind a health conversation a more useful home. This is planned scope, not an announcement that the Health space is already available.

For someone using it, the important details are ordinary ones. Which document am I looking at? Is this the latest version? Can I find the one I used last time? These are product questions we want to answer clearly, without asking people to learn a new filing system just to discuss their health.

There is also a difference between a document someone brings to a single conversation and information they want to keep using over time. The interface needs to make that choice understandable. People should know what they selected and what the product will do with it, without having to guess from a broad promise about file support.

Our aim is a health workspace that feels straightforward enough to return to. We are not presenting every file type, device connection or future feature as part of today's beta. As the experience develops, each release should explain the particular thing someone can now do and where to find it.

The reason for doing this work is simple. A person who has taken the trouble to collect information about their health should have an easier way to use it when a question comes up.

https://anistratenco.com

#healthai #productdesign #digitalhealth

Health Assistant: the workspace in the public beta

Anistratenco — Mon, 21 Sep 2026 14:04:44 +0000

A health question often starts somewhere other than a chat window. It might start in a lab report, an article saved on a phone or a note from an earlier appointment. By the time someone types a question, much of the information behind it is still scattered elsewhere.

Health Assistant is the first public-beta product from Anistratenco. Its current web workspace brings conversation, selected files and cited research together. The practical value of that combination is straightforward: the material someone wants to understand can sit beside the discussion about it.

The workspace comes first

A person looking through a document should not have to compress everything into one polished prompt before a conversation can begin. Selected files give the discussion something concrete to refer to. Follow-up messages can add a missing detail or explain why one part of the document matters more than another.

Citations provide another route back to the underlying material. They are not a guarantee that an answer is correct, but they make it possible to inspect a source instead of treating the summary as the end of the process. That is an important part of the experience we want to improve.

Continuity is the next direction

The longer-term ambition is a health workspace that remains useful as circumstances change. A goal set earlier in the year may no longer be relevant. A new document may add context to an old concern. Our direction is to make those changes easier to carry into future conversations, while leaving the person in control of the picture being built.

Deeper longitudinal context, proactive workflows, trends, check-ins, realtime voice, and mobile and watch experiences belong to the roadmap. They should not be read as a list of capabilities already delivered in this beta. We will describe availability as those releases become usable.

For now, the public release is focused on the web conversation and the information brought into it. We want that starting point to be clear, useful and worth returning to.

https://anistratenco.com

healthai #digitalhealth #productdesign

Trying Health Assistant with a deliberately incomplete question

Anistratenco — Sat, 19 Sep 2026 21:35:05 +0000

You can try Health Assistant without uploading a medical record. A small fictional question is enough to explore something important about the experience: what happens when the information you provide is incomplete, and what happens when you correct it?

Consider this example: “I found an old note that says my result was 12, but I cannot see the test name, units or date. Help me work out what information I need before interpreting it.”

The missing details are the point. There is no sound interpretation to offer from that number alone. For this task, a useful answer would identify what needs checking rather than guess which test you meant.

Try a correction in the same conversation

Open https://anistratenco.com and try the fictional prompt. Read the response, then add: “That number was copied incorrectly. I do not have the original report yet.” Look at whether the next answer acknowledges the correction and stops relying on the earlier number.

This is a practical way to explore the conversational scope of the beta. You are not testing how confidently an assistant can produce an explanation. You are checking whether the conversation remains useful when the starting point changes. That matters because real questions often begin with a half-remembered detail, an unfamiliar phrase or a document that is not yet at hand.

Help us improve a specific moment

If the answer misses your correction, tell us what you entered, which sentence was unhelpful and what you expected instead. A fictional example makes that feedback easier to share without disclosing your health history or someone else’s records. Please keep personal medical information out of public comments.

Our ambition for Anistratenco is a health workspace that remains useful as understanding develops. The first public beta gives us a place to work on that with people using it, one concrete task at a time. This exercise describes the behaviour we want you to evaluate, not a guarantee that every response already meets that standard.

Watch the introduction: https://www.youtube.com/watch?v=gJRlemOH9EE

Explore Health Assistant: https://anistratenco.com

#healthai #healthassistant #ai #feedback #healthliteracy

A clearer starting point for preventive health

Anistratenco — Sun, 13 Sep 2026 12:45:07 +0000

We are building Anistratenco Health Assistant around a practical question: can preventive health become easier to understand and act on when the relevant context is connected instead of scattered?

The first public beta, planned for September 15, is intentionally a beginning. It starts with the way people actually approach their health: a result they want to understand, a concern that has not become a clear question, a change in routine, or a source they want to relate to their own situation.

What the first release is for

The first release is a conversational workspace for turning that starting point into a clearer next step. The experience is built around a few concrete behaviors:

Bring a question in ordinary language, without having to translate it into a perfect prompt.
Keep the relevant context together so a result is not read as if it arrived in isolation.
Connect explanations to the material and evidence behind them.
Make uncertainty and open questions visible instead of smoothing them away.
Let the person correct an outdated detail or a misunderstanding so the conversation can move forward with a better picture.

These are not separate features for a checklist. Together, they describe the kind of product we want to build: something that helps people move from information to understanding, and from understanding to a more deliberate choice.

Why the context matters

Health changes with the person. A routine shifts, a priority moves, a new result arrives, or an earlier assumption turns out to be wrong. A useful health workspace should make those changes legible. It should help someone see what has changed, what the available evidence supports, and what still needs attention.

That is the foundation for what we call health alignment. It is not a slogan and it is not a promise of a particular outcome. It is a clearer relationship between what you know, what you value, what is changing, and what you choose to do next.

What comes after the beta

The September release is the first public expression of a larger Anistratenco direction: a more proactive, longitudinal approach to health intelligence. Later work will expand the ways people can revisit their context, notice meaningful changes, and stay connected to the decisions that matter to them. Those are plans, not claims about what is already available.

The feedback we want is concrete. Where did the product lose the thread of your question? Which explanation needed more context? What would have made the next step easier to consider? Please keep public feedback general and do not include private health records.

The goal is simple: give people a clearer way to understand their health over time, with the person still at the centre of the decisions.

Three everyday tasks we want Health Assistant to get right

Anistratenco — Wed, 09 Sep 2026 09:38:57 +0000

A feature list can tell you what a product contains without telling you whether you can finish the task you opened it for. For Anistratenco Health Assistant, we want to evaluate the first beta through a few ordinary situations rather than treating the presence of a chat box or an upload button as sufficient.

Follow a reference back to the document

Imagine using a sample report to ask what a particular term means. The intended experience is to keep the explanation connected to the relevant material, so you can distinguish what the report actually says from the assistant’s interpretation. A useful test is whether you can locate the passage being discussed without searching the entire document again.

This is an illustrative product scenario, not a patient case or a claim of clinical validation.

Keep two projects separate

Someone may be preparing questions for an appointment while also exploring a longer-term fitness goal. Grouping the conversations should make it possible to return to the right task without mixing the two into one indistinct history. We want to learn whether the workspace’s organization matches how people naturally separate those concerns.

Understand what happened when an action fails

If a document cannot be read, the experience should make that clear rather than leave you wondering whether its contents influenced the answer. Likewise, an unfinished action needs an understandable status and a sensible next step. A successful-looking screen is not enough if the person cannot tell what was actually completed.

What these scenarios mean for beta

These are criteria we are working toward, not a report that every scenario has passed or that every planned feature will be available at launch. The planned September 15, 2026 beta should give us a way to observe the gaps, explain the actual supported scope, and improve the tasks people struggle to finish.

Health Assistant’s consumer role is to help with understanding and organization; it does not turn these workflows into diagnosis, prescribing or emergency care.

Anistratenco Health Assistant: what we’re building for the public beta

Anistratenco — Tue, 08 Sep 2026 17:10:10 +0000

The first release and the longer-term roadmap serve different purposes, and we want that distinction to be clear before anyone signs up. This overview sets out the feature areas planned for Anistratenco Health Assistant’s September 15, 2026 public beta and the capabilities intended for later releases.

What we’re working on first

Conversations about your health

The starting point is a place to ask questions in ordinary language and follow up when something is unclear. We want you to be able to explore a topic, ask for a simpler explanation, or prepare questions for an appointment without starting from scratch each time.

Documents and results alongside the conversation

We’re working on bringing PDFs, images, and health documents into that same workspace. The intended use is practical: upload a report, ask about a term you don’t understand, or organize the points you want to discuss further. Keeping the original material close to the conversation makes it easier to check what an explanation refers to.

A personal health avatar

The health avatar is our name for an evolving picture of your health context: your goals, relevant history, measurements, and information you choose to share. We’re building it so that you can review and correct that picture as your circumstances change.

The purpose is continuity. If you come back to a question later, relevant background should not need to be reconstructed from the beginning.

Sources you can follow

We want explanations to connect back to supporting sources and make uncertainty understandable. Someone reading about a health topic should be able to explore the evidence behind an answer, not just receive a summary with nowhere to go next.

Organization and privacy controls

We’re also working on projects for grouping related conversations and goals, and controls over what personal context is retained or shared. These are part of making the workspace useful for ongoing questions rather than a single visit.

What comes after the beta

Our longer-term roadmap includes custom health trends, check-ins, reminders, and more proactive follow-up. We also plan to expand real-time voice and calling experiences, alongside dedicated iOS, Android, Apple Watch, and Android watch apps.

Those are later milestones, not a promise that every feature will be included on September 15. Our current target for the broader v1 release is January 15, 2027, with scope and timing subject to what we learn during the beta.

Who the first version is for

We’re starting with people who already take an active interest in their health and want a more organized way to understand it. Preparing for an appointment, revisiting an earlier result, or keeping track of questions around a personal goal are the kinds of everyday situations we’re designing for.

The initial scope is understanding and organizing health information, not replacing a clinician or providing emergency care.

We’ll share the confirmed beta feature list and access details with the September 15 release announcement. Feedback on what is useful, confusing, or missing will help shape what we build next.