<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Pinnacle Crypt (The Enchanter)</title>
    <description>The latest articles on DEV Community by Pinnacle Crypt (The Enchanter) (@pinnaclecrypt).</description>
    <link>https://dev.to/pinnaclecrypt</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4079622%2Fec92dc98-f37a-4130-ae3f-dc32899d0b64.png</url>
      <title>DEV Community: Pinnacle Crypt (The Enchanter)</title>
      <link>https://dev.to/pinnaclecrypt</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/pinnaclecrypt"/>
    <language>en</language>
    <item>
      <title>Anamnesis: a medical history that remembers for you</title>
      <dc:creator>Pinnacle Crypt (The Enchanter)</dc:creator>
      <pubDate>Thu, 05 Nov 2026 10:19:51 +0000</pubDate>
      <link>https://dev.to/pinnaclecrypt/anamnesis-a-medical-history-that-remembers-for-you-4bk9</link>
      <guid>https://dev.to/pinnaclecrypt/anamnesis-a-medical-history-that-remembers-for-you-4bk9</guid>
      <description>&lt;p&gt;Every new doctor starts with the same question: "So, what's your history?" Most of us answer from memory, badly. We forget when a dose changed or which pill caused the rash, and we can't remember what the last specialist said to watch for. &lt;em&gt;Anamnesis&lt;/em&gt;, the Greek word for recollection, is also the medical term for that history. I built Anamnesis for the Tatum × Walrus Memory "Total Recall" hackathon, so the patient holds that history and it gets more useful with every visit.&lt;/p&gt;

&lt;h2&gt;
  
  
  What it does
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Your own record.&lt;/strong&gt; Tapping "Start my record" creates a private record key that stays on your device, plus a record card (the key and a QR code) for opening it on another phone. Each key maps to its own Walrus Memory namespace, so a refresh or a visit a week later brings back your history and nobody else's. If you want the account itself to be yours, you can link your own Walrus Memory account instead.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Note.&lt;/strong&gt; After a visit, flare-up, prescription or reading, write or dictate what happened in plain words, or photograph the prescription, lab sheet or discharge letter. Anamnesis splits it into atomic facts (symptom, medication, reaction, diagnosis, clinician instruction, measurement, follow-up), each dated, and files every fact as its own memory.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Today.&lt;/strong&gt; The home screen is a snapshot, not a feed: what you take now, what's still open (for example "ECG due"), allergies, the last visit, and what to raise next time. Every line links to the memories behind it. The full history lives in a searchable timeline.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Brief.&lt;/strong&gt; Before an appointment, say what it's for. Anamnesis writes a one-page brief: what changed, current meds, patterns across visits, questions to ask, and flags to raise. Every line cites the memories it rests on; tap a citation to read them and check each one's Walrus blob on Sui.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Doctor links and an emergency card.&lt;/strong&gt; A signed link gives one doctor read access to your record for a day, a week or three months, and you can revoke it whenever you like. The doctor sees medicines over time as a dose chart, what changed, flags and your questions, each with its source. The doctor can mark a flag or question &lt;strong&gt;Addressed&lt;/strong&gt; or &lt;strong&gt;Keep watching&lt;/strong&gt;, and that verdict is filed into your record. The emergency card puts allergies, current medicines, conditions and a contact behind a QR code that responders can scan.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Proof on Sui.&lt;/strong&gt; A page shows, live from Sui mainnet, that the record is held on Walrus and how long its storage is paid for.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  How it learns over time
&lt;/h2&gt;

&lt;p&gt;The demo record is a migraine history, five months and six notes long, now over 60 memories on mainnet. Here's what Anamnesis said back when each note was filed:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;June, first visit:&lt;/strong&gt; "This is your first entry, so no historical patterns yet." It only files what you said.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;July:&lt;/strong&gt; the GP says daily ibuprofen causes rebound headaches. Anamnesis connects that to June: "the ibuprofen you've been taking most days since at least June ... may explain why frequency stayed so high."&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;August:&lt;/strong&gt; after two weeks on topiramate, headaches are "down to about 2 a week (a big drop from the 14 days recorded on 2026-07-24)". It knows the baseline because it recalled the July memory before filing.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;September:&lt;/strong&gt; three ibuprofen days in one week. It flags that this breaks "the 2-day limit Dr. Okafor set in July".&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;October, neurology:&lt;/strong&gt; it files the switch to propranolol as "the final step in a chain that began with the cognitive side effects first noted after the dose increase back in August."&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That works because intake recalls related history &lt;em&gt;before&lt;/em&gt; filing, and when a new note connects to the past it files a &lt;code&gt;pattern&lt;/code&gt; memory. Later recalls find the link directly instead of working it out again.&lt;/p&gt;

&lt;p&gt;The brief for the 8-week neurology follow-up then asks whether the ECG was done, whether the word-finding trouble has cleared since stopping topiramate, and whether ibuprofen is staying under the limit. It also flags watching for wheeze on propranolol, given the childhood asthma. Each line cites the memories behind it. Briefs are filed too, so the next one knows which questions were already asked.&lt;/p&gt;

&lt;p&gt;Doctors teach it as well. When a doctor marks "ankle swelling on amlodipine" as addressed, a &lt;code&gt;review&lt;/code&gt; memory is filed in the patient's namespace: who reviewed it, what, and the verdict. Today, briefs and the doctor view all recall reviews, so the next brief stops listing that item as open, while anything marked "keep watching" stays in view with the doctor's note attached. The record now learns from both sides of the appointment.&lt;/p&gt;

&lt;h2&gt;
  
  
  How it uses Walrus Memory
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;@mysten-incubation/memwal&lt;/code&gt; on the mainnet relayer. Every patient's record gets its own namespace, &lt;code&gt;p_&lt;/code&gt; followed by a hash of their record key. The sample patient sits in its own namespace and is read-only.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;rememberBulk&lt;/code&gt; + &lt;code&gt;waitForRememberJobs&lt;/code&gt; handle multi-fact intake. Each fact is tagged &lt;code&gt;[kind · date]&lt;/code&gt;, which lets Today, the timeline and the doctor view sort and filter.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;recall&lt;/code&gt; runs several targeted queries in parallel (purpose, meds, reactions, instructions, earlier briefs) and merges the results by blob ID. Every citation links to its Walrus blob.&lt;/li&gt;
&lt;li&gt;Sharing is memory too. Issuing and revoking a doctor or emergency link each files a &lt;code&gt;share&lt;/code&gt; memory, and every open checks the signed link against those memories. Revoking takes effect on the next open.&lt;/li&gt;
&lt;li&gt;Doctor verdicts are memory too: each Addressed / Keep watching click files a &lt;code&gt;review&lt;/code&gt; memory, scoped by the signed doctor link to that one patient's namespace.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;listMemories&lt;/code&gt; feeds Proof on Sui with blob IDs, sizes, and the epoch the storage is paid through.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  How it uses Tatum
&lt;/h2&gt;

&lt;p&gt;All chain reads go through Tatum's Sui mainnet gRPC gateway. Tatum marks Sui JSON-RPC as deprecated, so this uses the modern path: &lt;code&gt;SuiGrpcClient&lt;/code&gt; over native gRPC with &lt;code&gt;x-api-key&lt;/code&gt; metadata. One gotcha: the gateway speaks native gRPC, not gRPC-web, so you need &lt;code&gt;@protobuf-ts/grpc-transport&lt;/code&gt; with &lt;code&gt;@grpc/grpc-js&lt;/code&gt; rather than the default fetch transport.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Every citation is checked on Sui.&lt;/strong&gt; Each memory is a Walrus &lt;code&gt;Blob&lt;/code&gt; object owned by the account's wallet. When you open a citation, the server lists that wallet's &lt;code&gt;Blob&lt;/code&gt; objects through Tatum (&lt;code&gt;listOwnedObjects&lt;/code&gt; filtered to the Walrus blob type), converts each u256 &lt;code&gt;blob_id&lt;/code&gt; to its base64url form, and matches it to the cited memory. The popover shows the Sui object, the epoch it was certified in, the epoch its storage is paid through (with the approximate date) and the checkpoint it was read at. A cited line is then traceable from the sentence to the memory to the object on chain.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Proof on Sui&lt;/strong&gt; reads the MemWal &lt;code&gt;Account&lt;/code&gt; object (owner, version, registered delegate keys, last transaction), the Walrus system object's current epoch, and the latest checkpoint.&lt;/li&gt;
&lt;li&gt;The &lt;strong&gt;doctor view&lt;/strong&gt; puts the account proof under every page, and its citations get the same per-blob check: the Sui checkpoint and the Walrus epoch at the moment the doctor opened the link.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Your own account:&lt;/strong&gt; if a patient links their own Walrus account, a doctor can also use one of its delegate keys. Over Tatum, the server then checks that the key is still registered on the account object before it recalls anything.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Try it
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Demo: &lt;a href="https://anamnesis-inky.vercel.app" rel="noopener noreferrer"&gt;https://anamnesis-inky.vercel.app&lt;/a&gt; (start a record, or open the sample patient)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Code: &lt;a href="https://github.com/PinnacleCryptNG/anamnesis" rel="noopener noreferrer"&gt;https://github.com/PinnacleCryptNG/anamnesis&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Built with Next.js, Walrus Memory, Tatum, and Claude. Anamnesis organises what you tell it; it does not diagnose.&lt;/p&gt;

&lt;p&gt;&lt;a class="mentioned-user" href="https://dev.to/tatum_io"&gt;@tatum_io&lt;/a&gt; @WalrusFoundation&lt;/p&gt;

</description>
      <category>hackathon</category>
      <category>privacy</category>
      <category>softwaredevelopment</category>
    </item>
    <item>
      <title>Handover: I made three AI agents run an exit interview in team chat</title>
      <dc:creator>Pinnacle Crypt (The Enchanter)</dc:creator>
      <pubDate>Wed, 04 Nov 2026 16:20:08 +0000</pubDate>
      <link>https://dev.to/pinnaclecrypt/handover-i-made-three-ai-agents-run-an-exit-interview-in-team-chat-57mf</link>
      <guid>https://dev.to/pinnaclecrypt/handover-i-made-three-ai-agents-run-an-exit-interview-in-team-chat-57mf</guid>
      <description>&lt;p&gt;When someone leaves a team, the things they know usually leave with them. You find out&lt;br&gt;
three weeks later, when a cron job fails at 2 a.m. and the only person who knew the fix is&lt;br&gt;
gone. The runbook says "TODO".&lt;/p&gt;

&lt;p&gt;For the &lt;a href="https://www.wemakedevs.org/hackathons/switch" rel="noopener noreferrer"&gt;Flip the Switch hackathon&lt;/a&gt; I built&lt;br&gt;
&lt;strong&gt;Handover&lt;/strong&gt;. Three AI agents and the people involved run the handover in an ordinary chat&lt;br&gt;
channel. &lt;/p&gt;

&lt;p&gt;This post covers what it does, how to set it up from zero, and what I got wrong&lt;br&gt;
along the way.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Repo:&lt;br&gt;
&lt;a href="https://github.com/PinnacleCryptNG/handover" rel="noopener noreferrer"&gt;https://github.com/PinnacleCryptNG/handover&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Demo video:&lt;br&gt;
&lt;a href="https://youtu.be/yEJa8bV6z0M?si=0Y18zIxiUrBnbWsD" rel="noopener noreferrer"&gt;https://youtu.be/yEJa8bV6z0M?si=0Y18zIxiUrBnbWsD&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;h2&gt;
  
  
  What it does
&lt;/h2&gt;

&lt;p&gt;Handover has two rooms and three agents:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Room&lt;/th&gt;
&lt;th&gt;People&lt;/th&gt;
&lt;th&gt;Agents&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;payments-handover&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Ada (leaving), Maya (manager)&lt;/td&gt;
&lt;td&gt;Interviewer, Scribe, Buddy&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;payments-onboarding&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;Ben (taking over), Maya&lt;/td&gt;
&lt;td&gt;Buddy, Scribe&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;ol&gt;
&lt;li&gt;The &lt;strong&gt;Interviewer&lt;/strong&gt; reads the system's folder first: code, cron file, runbook. Then it
interviews Ada one question at a time. It asks what she owns, the recurring jobs, how
things break, access and contacts, open promises, and the "never do X" rules.&lt;/li&gt;
&lt;li&gt;The &lt;strong&gt;Scribe&lt;/strong&gt; checks each answer against the files. It writes &lt;code&gt;HANDOVER.md&lt;/code&gt;, where every
fact has a source: "(Ada, interview)" or "(ops/crontab:3)". Anything nobody confirmed
goes under &lt;em&gt;Open questions&lt;/em&gt;. Nothing is guessed.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Maya&lt;/strong&gt;, the manager, reads the guide in chat, asks for changes and replies "approve".&lt;/li&gt;
&lt;li&gt;Only then does the Scribe publish the guide as a room document in Ben's room.&lt;/li&gt;
&lt;li&gt;The &lt;strong&gt;Buddy&lt;/strong&gt; answers Ben's questions from the guide. If the guide has no answer, it asks
Ada in the other room, brings her reply back to Ben and tells the Scribe to update the
guide.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The sample system is a small, made-up payments sync (&lt;code&gt;workspace/ledger-sync&lt;/code&gt;). It has&lt;br&gt;
the kind of gaps real systems have: a month-end job the runbook marks TODO, a hard-coded&lt;br&gt;
FX fallback rate and a rate-limit workaround nobody wrote down.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fh1btyh1jn0wnolfsvb5d.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fh1btyh1jn0wnolfsvb5d.jpg" alt="Ada answers the Interviewer; Ben's room waits for the approved guide" width="799" height="422"&gt;&lt;/a&gt;&lt;/p&gt;
&lt;h2&gt;
  
  
  Why Switch
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://github.com/sandbox-quantum/switch" rel="noopener noreferrer"&gt;Switch&lt;/a&gt; connects coding agents like Claude Code,&lt;br&gt;
Codex and OpenCode to normal chat rooms. Five Switch features carry this project:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Templates&lt;/strong&gt;: one YAML file creates all three agents, both rooms, a room group, the links
between the rooms and the first message in each room.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;@mentions&lt;/strong&gt;: people and agents talk by mentioning each other in a normal Mattermost
channel. Slack, Teams, Discord and Telegram work the same way.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Linked rooms&lt;/strong&gt;: the Buddy can carry a question from Ben's room to Ada's room and back.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Room documents&lt;/strong&gt;: the approved guide is a document attached to the onboarding room,
so everyone there reads the same text.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Working directories&lt;/strong&gt;: all three agents run inside the system's folder, so the Scribe
can cite a file and line instead of trusting the interview.&lt;/li&gt;
&lt;/ul&gt;
&lt;h2&gt;
  
  
  Setting it up from zero
&lt;/h2&gt;

&lt;p&gt;You need Docker and an agent provider.&lt;/p&gt;
&lt;h3&gt;
  
  
  1. Switch Console and a local server
&lt;/h3&gt;

&lt;p&gt;Download Switch Console from the&lt;br&gt;
&lt;a href="https://github.com/sandbox-quantum/switch/releases" rel="noopener noreferrer"&gt;releases page&lt;/a&gt; and choose&lt;br&gt;
&lt;strong&gt;Run a local server&lt;/strong&gt;. Console starts the Switch server and a Mattermost instance in Docker&lt;br&gt;
on your machine, so you don't need a Slack workspace to try it.&lt;/p&gt;

&lt;p&gt;Two Linux tips that cost me time:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Mattermost might not be on port 8065.&lt;/strong&gt; Console picks free ports. Read the real one from
&lt;code&gt;~/.config/switchdash/local-switch-server/ports.json&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Secret storage.&lt;/strong&gt; On a Linux machine without a desktop keyring, Console warns that secure
storage is unavailable. Installing &lt;code&gt;gnome-keyring&lt;/code&gt; and starting Console with
&lt;code&gt;--password-store=gnome-libsecret&lt;/code&gt; fixed it.&lt;/li&gt;
&lt;/ul&gt;
&lt;h3&gt;
  
  
  2. An agent provider
&lt;/h3&gt;

&lt;p&gt;I used &lt;a href="https://opencode.ai" rel="noopener noreferrer"&gt;OpenCode&lt;/a&gt; with Kimi K2.6 on Nebius Token Factory. Any&lt;br&gt;
OpenAI-compatible endpoint works the same way. This is &lt;code&gt;~/.config/opencode/opencode.json&lt;/code&gt;:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"provider"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="nl"&gt;"nebius"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"npm"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"@ai-sdk/openai-compatible"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"options"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt;
        &lt;/span&gt;&lt;span class="nl"&gt;"baseURL"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"https://api.tokenfactory.nebius.com/v1/"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt;
        &lt;/span&gt;&lt;span class="nl"&gt;"apiKey"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"{env:OPENAI_API_KEY}"&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="p"&gt;},&lt;/span&gt;&lt;span class="w"&gt;
      &lt;/span&gt;&lt;span class="nl"&gt;"models"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"moonshotai/Kimi-K2.6"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"name"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"Kimi K2.6"&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
    &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="p"&gt;},&lt;/span&gt;&lt;span class="w"&gt;
  &lt;/span&gt;&lt;span class="nl"&gt;"model"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"nebius/moonshotai/Kimi-K2.6"&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Export your key as &lt;code&gt;OPENAI_API_KEY&lt;/code&gt; before starting Console. The key stays in your&lt;br&gt;
environment and never goes into the file. Claude Code or Codex also work: pick them under&lt;br&gt;
&lt;em&gt;Advanced&lt;/em&gt; when you use the template.&lt;/p&gt;
&lt;h3&gt;
  
  
  3. Users
&lt;/h3&gt;

&lt;p&gt;You need one person for each role. For a demo, create them in the local Mattermost:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="nv"&gt;C&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="si"&gt;$(&lt;/span&gt;docker ps &lt;span class="nt"&gt;--format&lt;/span&gt; &lt;span class="s1"&gt;'{{.Names}}'&lt;/span&gt; | &lt;span class="nb"&gt;grep &lt;/span&gt;mattermost&lt;span class="si"&gt;)&lt;/span&gt;
&lt;span class="k"&gt;for &lt;/span&gt;u &lt;span class="k"&gt;in &lt;/span&gt;ada ben maya&lt;span class="p"&gt;;&lt;/span&gt; &lt;span class="k"&gt;do
  &lt;/span&gt;docker &lt;span class="nb"&gt;exec&lt;/span&gt; &lt;span class="nv"&gt;$C&lt;/span&gt; mmctl &lt;span class="nt"&gt;--local&lt;/span&gt; user create &lt;span class="nt"&gt;--username&lt;/span&gt; &lt;span class="nv"&gt;$u&lt;/span&gt; &lt;span class="nt"&gt;--email&lt;/span&gt; &lt;span class="nv"&gt;$u&lt;/span&gt;@example.com &lt;span class="nt"&gt;--password&lt;/span&gt; &lt;span class="s2"&gt;"&amp;lt;choose one&amp;gt;"&lt;/span&gt;
  docker &lt;span class="nb"&gt;exec&lt;/span&gt; &lt;span class="nv"&gt;$C&lt;/span&gt; mmctl &lt;span class="nt"&gt;--local&lt;/span&gt; team &lt;span class="nb"&gt;users &lt;/span&gt;add switch &lt;span class="nv"&gt;$u&lt;/span&gt;
&lt;span class="k"&gt;done&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  4. The template
&lt;/h3&gt;

&lt;p&gt;In Console, open &lt;strong&gt;Templates → Import template&lt;/strong&gt;, pick&lt;br&gt;
&lt;a href="https://github.com/PinnacleCryptNG/handover/blob/main/templates/handover.template.yaml" rel="noopener noreferrer"&gt;&lt;code&gt;templates/handover.template.yaml&lt;/code&gt;&lt;/a&gt;,&lt;br&gt;
then &lt;strong&gt;Use template&lt;/strong&gt;. The form comes from the template's &lt;code&gt;params&lt;/code&gt;: a name for the&lt;br&gt;
handover, the three people, the system folder and the chat app. The preview on the right&lt;br&gt;
shows what will be created.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ffvq9yxyvvwl5zg0k229c.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ffvq9yxyvvwl5zg0k229c.jpg" alt="The template form and its preview: three agents, two rooms" width="799" height="422"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The rooms and links part of the template is short:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight yaml"&gt;&lt;code&gt;&lt;span class="na"&gt;rooms&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt;
  &lt;span class="pi"&gt;-&lt;/span&gt; &lt;span class="na"&gt;name&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-handover"&lt;/span&gt;
    &lt;span class="na"&gt;agents&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="pi"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-interviewer"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-scribe"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-buddy"&lt;/span&gt;&lt;span class="pi"&gt;]&lt;/span&gt;
    &lt;span class="na"&gt;users&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="pi"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{$creator}"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{leaver}"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{manager}"&lt;/span&gt;&lt;span class="pi"&gt;]&lt;/span&gt;
    &lt;span class="na"&gt;kickoff&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="pi"&gt;|&lt;/span&gt;
      &lt;span class="s"&gt;@{slug}-interviewer @{leaver} is leaving and hands {slug} over to&lt;/span&gt;
      &lt;span class="s"&gt;@{successor}. Read the system folder, then greet @{leaver} in one line&lt;/span&gt;
      &lt;span class="s"&gt;and ask your first question.&lt;/span&gt;
  &lt;span class="pi"&gt;-&lt;/span&gt; &lt;span class="na"&gt;name&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-onboarding"&lt;/span&gt;
    &lt;span class="na"&gt;agents&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="pi"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-buddy"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-scribe"&lt;/span&gt;&lt;span class="pi"&gt;]&lt;/span&gt;
    &lt;span class="na"&gt;users&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="pi"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{$creator}"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{successor}"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{manager}"&lt;/span&gt;&lt;span class="pi"&gt;]&lt;/span&gt;

&lt;span class="na"&gt;links&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt;
  &lt;span class="pi"&gt;-&lt;/span&gt; &lt;span class="pi"&gt;{&lt;/span&gt; &lt;span class="nv"&gt;from&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-onboarding"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="nv"&gt;to&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-handover"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="nv"&gt;label&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="nv"&gt;interview with the leaver&lt;/span&gt; &lt;span class="pi"&gt;}&lt;/span&gt;
  &lt;span class="pi"&gt;-&lt;/span&gt; &lt;span class="pi"&gt;{&lt;/span&gt; &lt;span class="nv"&gt;from&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-handover"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="nv"&gt;to&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"&lt;/span&gt;&lt;span class="s"&gt;{slug}-onboarding"&lt;/span&gt;&lt;span class="pi"&gt;,&lt;/span&gt; &lt;span class="nv"&gt;label&lt;/span&gt;&lt;span class="pi"&gt;:&lt;/span&gt; &lt;span class="nv"&gt;successor's room&lt;/span&gt; &lt;span class="pi"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Most of the work is in each agent's &lt;code&gt;instructions&lt;/code&gt;, which are plain English. Click&lt;br&gt;
&lt;strong&gt;Create all&lt;/strong&gt;, and the Interviewer greets the leaver within seconds.&lt;/p&gt;

&lt;h3&gt;
  
  
  5. Run it
&lt;/h3&gt;

&lt;p&gt;Sign in to Mattermost as &lt;code&gt;ada&lt;/code&gt; and answer the questions. Sign in as &lt;code&gt;maya&lt;/code&gt; to review the&lt;br&gt;
guide, and as &lt;code&gt;ben&lt;/code&gt; to ask the Buddy things. If you want to replay a run,&lt;br&gt;
&lt;a href="https://github.com/PinnacleCryptNG/handover/blob/main/scripts/mm.py" rel="noopener noreferrer"&gt;&lt;code&gt;scripts/mm.py&lt;/code&gt;&lt;/a&gt;&lt;br&gt;
posts and reads messages as any user through the Mattermost API.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0by5zj8fxvcpymgcju5b.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0by5zj8fxvcpymgcju5b.jpg" alt="The Scribe's summary, Maya's change request and her approval" width="799" height="422"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fapb6p7xgyhez2gysd72w.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fapb6p7xgyhez2gysd72w.jpg" alt="Ben asks two questions; the Buddy answers one from the guide and gets the other from Ada" width="799" height="422"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I learned
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;An agent only sees messages that mention it.&lt;/strong&gt; This caught me during the recorded demo.&lt;br&gt;
The Interviewer posted its notes as plain messages, without &lt;code&gt;@payments-scribe&lt;/code&gt; in front.&lt;br&gt;
It finished with "interview complete", and the Scribe never woke up. The instructions said&lt;br&gt;
"tell the scribe", which isn't the same as "start the message with @payments-scribe". The&lt;br&gt;
fix was to spell it out and give the reason: &lt;em&gt;"post a note that starts with&lt;br&gt;
@{slug}-scribe (the scribe only sees messages that mention it)"&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Name the role, not just the person.&lt;/strong&gt; In one run, the Scribe asked Ada to approve her own&lt;br&gt;
guide, probably because she was the last person who had talked to it. Changing "ask @{manager}" to&lt;br&gt;
"ask @{manager} (the manager, not the leaver)" fixed it. With templates, a placeholder&lt;br&gt;
becomes a username at runtime, so the agent can lose track of what that person's role is.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;"Read the folder first" is the most useful line in the template.&lt;/strong&gt; Because the&lt;br&gt;
Interviewer reads the code before it asks anything, its first question already names the&lt;br&gt;
system: "You own ledger-sync — who else depends on it staying healthy?" The Scribe also&lt;br&gt;
caught a gap Ada never mentioned: the month-end section of the runbook still says TODO.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Make the agent cite sources.&lt;/strong&gt; Requiring a source on every fact, and an &lt;em&gt;Open questions&lt;/em&gt;&lt;br&gt;
section for everything else, made the guide trustworthy. It also gave the Buddy a clear&lt;br&gt;
rule: if the answer isn't in the guide, don't make one up, ask Ada.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Approval should be a real gate.&lt;/strong&gt; The Scribe publishes only after the manager replies&lt;br&gt;
"approve" in chat. Maya's change ("make Ben the owner of the token rotation") shows up&lt;br&gt;
in the guide before it reaches Ben.&lt;/p&gt;

&lt;h2&gt;
  
  
  Try it
&lt;/h2&gt;

&lt;p&gt;Clone the repo, import the template and point &lt;em&gt;System folder&lt;/em&gt; at &lt;code&gt;workspace/ledger-sync&lt;/code&gt;,&lt;br&gt;
or at a real system you're about to hand over. The full transcript of a run is in&lt;br&gt;
&lt;a href="https://github.com/PinnacleCryptNG/handover/blob/main/examples/transcript.md" rel="noopener noreferrer"&gt;&lt;code&gt;examples/transcript.md&lt;/code&gt;&lt;/a&gt;&lt;br&gt;
and the guide it produced is in&lt;br&gt;
&lt;a href="https://github.com/PinnacleCryptNG/handover/blob/main/examples/HANDOVER.md" rel="noopener noreferrer"&gt;&lt;code&gt;examples/HANDOVER.md&lt;/code&gt;&lt;/a&gt;.&lt;/p&gt;

&lt;h1&gt;
  
  
  HowISwitch
&lt;/h1&gt;

</description>
      <category>ai</category>
      <category>webdev</category>
      <category>programming</category>
      <category>javascript</category>
    </item>
    <item>
      <title>TAILORY 🐾 — Google Gemini AI Dog Day 2026 Storycraft &amp; Solana Devnet Keepsake Engine</title>
      <dc:creator>Pinnacle Crypt (The Enchanter)</dc:creator>
      <pubDate>Sun, 16 Aug 2026 12:03:11 +0000</pubDate>
      <link>https://dev.to/pinnaclecrypt/tailory-google-gemini-ai-dog-day-2026-storycraft-solana-devnet-keepsake-engine-1d71</link>
      <guid>https://dev.to/pinnaclecrypt/tailory-google-gemini-ai-dog-day-2026-storycraft-solana-devnet-keepsake-engine-1d71</guid>
      <description>&lt;p&gt;&lt;strong&gt;This is my submission for the &lt;a href="https://dev.to/challenges/weekend-2026-08-15"&gt;Weekend Challenge: Dog Days Edition&lt;/a&gt;.&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;Most modern pet applications treat our dogs as cold analytical metrics: weight charts, calorie counters, and step trackers. &lt;/p&gt;

&lt;p&gt;But on &lt;strong&gt;International Dog Day 2026&lt;/strong&gt;, what we truly treasure isn't a spreadsheet it's their bizarre daily rituals, the warmth of their silent companionship, the shoes they've stolen, and the unspoken bond that makes them family.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;TAILORY&lt;/strong&gt; is an editorial storycraft and memory preservation companion that turns your dog's unique personality and cherished memories into an unforgettable, multimodal keepsake.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fquotbluwpzz96yu69tsy.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fquotbluwpzz96yu69tsy.png" alt="TAILORY banner showcasing Google Gemini AI canine character analysis, spoken voice companion, and Dog Day 2026 Solana keepsakes" width="800" height="450"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Watch the complete walkthrough on Youtube: TAILORY Video Demo&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;  &lt;iframe src="https://www.youtube.com/embed/RGopewtv8AA" width="710" height="399"&gt;
  &lt;/iframe&gt;
&lt;/p&gt;

&lt;h2&gt;
  
  
  Live Demo &amp;amp; Repository
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Live Interactive Application&lt;/strong&gt;: &lt;a href="https://ais-pre-vfnndg2hhbm7prqkr3uvp3-123668770786.europe-west2.run.app" rel="noopener noreferrer"&gt;Launch TAILORY&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Google AI Studio Applet&lt;/strong&gt;: &lt;a href="https://ai.studio/apps/34524fa1-35b6-4687-b783-bd7abd5920c0" rel="noopener noreferrer"&gt;Open in AI Studio&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Key Features &amp;amp; Architecture
&lt;/h2&gt;

&lt;h3&gt;
  
  
  1.  Multimodal Canine Character Blueprint (Google Gemini 3.7 Flash)
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Upload your dog's photo or pick from instant archetypes (e.g., &lt;em&gt;Bruno the Snack Strategist&lt;/em&gt;, &lt;em&gt;Luna the Velvet Goblin&lt;/em&gt;, &lt;em&gt;Barnaby the Sofa Philosopher&lt;/em&gt;).&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Gemini analyzes 5 expressive behavioral dimensions—happiest tail-wag triggers, favorite sneaker-theft habits, and social quirks—to construct a rich psychological character profile.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  2.  Spoken Canine Companion (ElevenLabs &amp;amp; React Audio Waves)
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Talk directly to your dog in real-time. Ask them why they stole your socks, what they dream about during twitchy naps, or who their favorite human is.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Features dynamic voice synthesis and reacting soundwave visualizers for an expressive conversational experience.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  3.  Heartfelt Storycraft Narrative Engine
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Weaves 4 real memories (&lt;em&gt;How We Met&lt;/em&gt;, &lt;em&gt;Unforgettable Adventures&lt;/em&gt;, &lt;em&gt;Funniest Habits&lt;/em&gt;, and &lt;em&gt;Heartfelt Love Message&lt;/em&gt;) into a 3-act literary keepsake story.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Formatted with editorial drop-caps, pull quotes, and a built-in read-aloud narrator.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  4.  Official Keepsake Certificate &amp;amp; High-Res PNG Export
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Generates a certificate featuring your dog’s official title, character traits, and literary excerpt.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;1-click 800×1100 high-resolution PNG export suitable for physical framing and printing.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  5.  Solana Devnet On-Chain Memory Preservation
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Computes a SHA-256 cryptographic memory digest of your dog’s story.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Records the metadata on the &lt;strong&gt;Solana Devnet blockchain&lt;/strong&gt; using the SPL Memo standard for verifiable on-chain permanence with live Solana Explorer transaction links.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;Tech Stack&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Frontend&lt;/strong&gt;: React 18, TypeScript, Tailwind CSS, Motion (framer-motion)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;AI &amp;amp; Multimodal Intelligence&lt;/strong&gt;: Google Gemini 3.7 Flash API (&lt;code&gt;@google/genai&lt;/code&gt;)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Voice &amp;amp; Audio&lt;/strong&gt;: ElevenLabs Voice Synthesis &amp;amp; Web Audio API&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;On-Chain Preservation&lt;/strong&gt;: Solana Web3.js (&lt;code&gt;@solana/web3.js&lt;/code&gt;), Solana Devnet SPL Memo&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Graphics &amp;amp; Export&lt;/strong&gt;: HTML5 Canvas Rendering Engine (2D Context)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Github Repo&lt;/strong&gt;: &lt;a href="https://github.com/PinnacleCryptNG/Tailory" rel="noopener noreferrer"&gt;https://github.com/PinnacleCryptNG/Tailory&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;h3&gt;
  
  
  1. Google Gemini AI Multimodal Personality &amp;amp; Story Engine
&lt;/h3&gt;

&lt;p&gt;We integrated the modern &lt;code&gt;@google/genai&lt;/code&gt; TypeScript SDK with &lt;code&gt;gemini-3.7-flash&lt;/code&gt;. The server-side API receives the dog's photo and user answers to perform multimodal reasoning:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;It constructs a custom character profile, archetype, and daily antics while adhering strictly to entertainment and emotional keepsake safety rules (no clinical or veterinary claims).&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;For the Storycraft stage, Gemini weaves 4 distinct memory prompts into an emotionally resonant 3-act narrative with quotes and dedications.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  2. ElevenLabs Expressive Spoken Voice
&lt;/h3&gt;

&lt;p&gt;To give the dog a true voice, we integrated the ElevenLabs Text-to-Speech API. &lt;/p&gt;

&lt;p&gt;Dog responses are synthesized with natural cadence and expressiveness settings. &lt;/p&gt;

&lt;p&gt;We also built an intelligent Web Speech API fallback with Web Audio API chime sound effects so the audio experience works seamlessly in every browser environment.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. Solana Devnet Memory Preservation
&lt;/h3&gt;

&lt;p&gt;To ensure memories last forever, we created an on-chain digital preservation layer:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;The app computes a SHA-256 cryptographic hash of the keepsake certificate, dog profile, and story text.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;It records a transaction on Solana Devnet, generating a verifiable transaction signature, slot number, and direct link to the Solana Explorer.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  4. Natural Editorial Design System
&lt;/h3&gt;

&lt;p&gt;We rejected generic dark-mode templates in favor of a warm, comforting storybook aesthetic:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Palette&lt;/strong&gt;: Forest Green (&lt;code&gt;#1b382b&lt;/code&gt;), Golden Amber (&lt;code&gt;#e5a93c&lt;/code&gt;), Warm Cream (&lt;code&gt;#faf6ee&lt;/code&gt;), and Earth Tones.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Typography &amp;amp; Motion&lt;/strong&gt;: Editorial serif display typography paired with organic fluid transitions via &lt;code&gt;motion/react&lt;/code&gt;.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

&lt;p&gt;We are submitting TAILORY for the following prize categories:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Best Use of Google AI&lt;/strong&gt;: Utilizes &lt;code&gt;@google/genai&lt;/code&gt; (&lt;code&gt;gemini-3.7-flash&lt;/code&gt;) for multimodal image analysis, dynamic canine chat persona generation, and narrative storycrafting.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Best Use of ElevenLabs&lt;/strong&gt;: Implements ElevenLabs expressive TTS voice synthesis to bring the dog companion to life with realistic spoken audio.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;&lt;strong&gt;Best Use of Solana&lt;/strong&gt;: Integrates Solana Devnet to permanently preserve dog memories with cryptographic SHA-256 verification and on-chain transaction explorer receipts.&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;em&gt;Dedicated to every good dog who made ordinary days our favourite days.&lt;/em&gt; 🐾&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;About the Developer&lt;/strong&gt;&lt;br&gt;
Hi, I'm Pinnacle! &lt;/p&gt;

&lt;p&gt;An aspiring developer and proud Google Developer Program member exploring the world of software and generative AI.&lt;/p&gt;

&lt;p&gt;As a beginner eager to learn by building, I recently created TAILORY 🐾—an AI dog storycraft app powered by Google Gemini and Solana Devnet for DEV.to’s Dog Days Challenge.&lt;/p&gt;

&lt;p&gt;I'm learning in public, building cool projects, and growing every day!&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>weekendchallenge</category>
      <category>ai</category>
      <category>gemini</category>
    </item>
  </channel>
</rss>
