mirror of
https://github.com/supermemoryai/supermemory.git
synced 2026-08-28 05:25:33 +00:00
434c86e2c3
Keep docs asset URLs repo-root-relative and pass images as literal MDX children so Mintlify can compile them through OptimizedImage under the /docs mount. Mintlify's official guidance says: “Image paths are root-relative from your docs repository.” It also says relative paths such as `./screenshot.png` are unsupported. See [Image embeds](https://www.mintlify.com/docs/create/image-embeds). The existing `/images/...` paths were correct. The failure came from passing them through custom-component string props, which kept Mintlify from seeing those images during its MDX transform. - fix hero, building-block, and Slack avatar images - preserve Slack avatar clipping - normalize Hermes and Company Brain icon sizing Tested with Mintlify validation, broken-link checks, and browser checks across every changed route.
49 lines
2.6 KiB
Text
49 lines
2.6 KiB
Text
---
|
|
title: "Long-Horizon Research"
|
|
sidebarTitle: "Long-Horizon Research"
|
|
description: "Questions that need many sources synthesized into one answer"
|
|
icon: "telescope"
|
|
---
|
|
|
|
import { SlackThread, SlackMessage, Mention } from "/snippets/slack-message.mdx";
|
|
|
|
## The scenario
|
|
|
|
Some questions aren't a quick lookup. They need pulling information from a lot of distinct places — docs, tickets, meetings, code, past Slack threads — and synthesizing it into one coherent answer. Closer to a research project than "what's the pricing?"
|
|
|
|
## What it will look like
|
|
|
|
<SlackThread channel="#product">
|
|
<SlackMessage self hasAvatar time="11:10 AM">
|
|
<img src="/images/company-brain/dhravya-slack-icon.jpg" alt="" />
|
|
<Mention>supermemory</Mention> prepare a brief on how we've handled enterprise SSO requests over the last two quarters — who asked, what we promised, what's still open, and what engineering thinks the real effort is
|
|
</SlackMessage>
|
|
<SlackMessage bot hasAvatar time="11:14 AM">
|
|
<img src="/images/company-brain/supermemory-slack-icon.png" alt="" />
|
|
**Enterprise SSO — two-quarter brief**
|
|
|
|
- **Demand:** 6 serious asks (Acme, Globex, Initech, …). Three converted to paid pilots.
|
|
- **Promises made:** Okta + Azure AD for pilots; custom SAML called "roadmap" in two decks.
|
|
- **Still open:** Globex blocked on SCIM; Initech waiting on audit-log export.
|
|
- **Eng view:** ~6 weeks for Okta/Azure path (from `#eng` thread Mar 18 + Linear ENG-610); SCIM is the long pole.
|
|
|
|
Sources: Q1/Q2 enterprise notes, Linear ENG-610, Granola "SSO planning", `#sales` threads. Want this as a Linear doc?
|
|
</SlackMessage>
|
|
</SlackThread>
|
|
|
|
## What's really going on
|
|
|
|
Today, a single turn can already hit multiple [connectors](/company-brain/connectors) and channel memory in one answer. Long-horizon research is for the cases where that isn't enough — the work needs more time, more parallel digging, and a structured synthesis step before it replies.
|
|
|
|
We're deliberately not documenting an internal "spawn N children" mechanism here. What matters for you: ask a research-shaped question, get a sourced brief back, without you having to run six separate searches.
|
|
|
|
Until it ships, break big questions into smaller ones (docs, then tickets, then "summarize what we have") — Company Brain already handles each of those well.
|
|
|
|
<CardGroup cols={2}>
|
|
<Card title="From docs" icon="book" href="/company-brain/use-cases/knowledge-recall">
|
|
What single-turn knowledge recall looks like today.
|
|
</Card>
|
|
<Card title="Meeting recall" icon="calendar-check" href="/company-brain/use-cases/meeting-recall">
|
|
Pulling decisions out of synced notes.
|
|
</Card>
|
|
</CardGroup>
|