{% extends "base.html" %}
{% block content %}
<div class="header">
  <div>
    <div class="kicker">Ecosystem</div>
    <h2 style="margin:8px 0 0">Use Your Memory Everywhere</h2>
    <div class="small">8mem keeps preferences and corrections portable instead of trapping them inside one assistant.</div>
  </div>
  <span class="badge gold">Portable memory layer</span>
</div>

<div class="grid cols-3">
  <div class="card"><h3>Browser UI</h3><p>Save, inspect, test, import, and edit memory locally before wiring any external agent.</p></div>
  <div class="card"><h3>Telegram bot</h3><p>Use memory from chat once BotFather and a public HTTPS webhook are configured.</p></div>
  <div class="card"><h3>Local models</h3><p>Ask with your saved memory using the Ollama model you choose during setup.</p></div>
  <div class="card"><h3>OpenClaw agent</h3><p>Connect Viri-style agent workflows after OpenClaw is installed on the machine.</p></div>
  <div class="card"><h3>Hermes agent</h3><p>Connect Govi-style agent workflows after Hermes is installed and ready.</p></div>
  <div class="card"><h3>Portable context</h3><p>Export a readable context block for ChatGPT, Claude, IDE assistants, or another tool.</p></div>
</div>

<div class="footer"><strong>8mem does not force one model.</strong> It stores memory locally and lets you reuse it with the agents and runtimes you choose.</div>
{% endblock %}