Apollo Space — o sistema operacional de IA da sua empresa

The last time someone asked “where’s the current pricing one-pager?”, the answer was three Slack threads, a stale Google Doc, and a PDF buried in an inbox. In Apollo, the answer is one document — and it’s the same one Athena reads before she drafts the proposal, and the same one Marcus reads to match your voice when he writes outbound. That’s the point: your company’s knowledge lives in one place, and people and agents work from the same source of truth. No more reconciling two answers.

What they’re for

Documents are rich-text pages where operational knowledge lives — the kind that doesn’t fit in a board card or a CRM lead:

Client briefings
Standard operating procedures (SOPs)
Decisions and meeting notes
Draft proposals
Voice + communication tone
Market research, analyses
Templates to clone

Unlike a file in Notion or Google Docs, these documents live inside the org’s context — agents can query them, leads can be mentioned, integrations can inject content. The Brain uses documents as one of its primary knowledge sources.

Hybrid mode — markdown + WYSIWYG

The editor is hybrid:

You type in markdown when you know it (shortcut)
Text renders as WYSIWYG automatically
Formatting buttons in the top bar work for those who prefer to click

Result: nobody needs to memorize markdown, but those who know it can type faster. # becomes H1, ## H2, - becomes a bullet, ` becomes inline code, etc.

What you can put in a doc

Rich text

Headers, bullets, numbered lists, blockquotes, code blocks with syntax highlighting (Python / JS / SQL / etc.), inline code.

Tables

Add column, add row, drag to reorder. No merged cells (keeps things simple).

Images

Direct upload (drag-drop) or paste from clipboard. Inline or block (width is your choice).

Embeds

External links become rich previews (with title + description + favicon). YouTube / Loom embed the player inline. PDFs become a viewer.

Mentions

@user notifies the person. @agent brings an agent into the doc (e.g., @scout review this briefing). #lead-uuid links to a specific CRM lead.

Checklist

Same syntax as markdown (- [ ]). Checking/unchecking persists to the whole doc — great for meeting notes with action items.

Collaboration

Simultaneous editing

Multiple people editing at the same time. Each person’s cursor appears with a color + name. Conflicts are resolved via CRDT — you don’t lose text even with poor latency.

Inline comments

Select a passage → the comment icon appears. Thread in the sidebar. Mention @user to notify. Comments can be resolved (they disappear) or left open for revisiting.

Version history

Each save generates a version. Open “History” to see a timeline + restore an older version. Versions are append-only — nothing is deleted.

Sharing

Agents and documents

Because documents live inside the Brain, agents consult them automatically during operations:

Marcus loads the org’s voice doc when composing outbound messages
Scout cross-references public research with internal docs (ICP, persona)
Athena reads a client briefing before proposing next steps

You can also invoke an agent directly in the doc via mention @scout or @athena — they respond inline with a comment or append at the end.

Templates

Templates are special documents that serve as blueprints:

Template	Purpose
Meeting notes	Name / date / attendees / decisions / action items
Client briefing	Company / pain / goal / context / next steps
Commercial proposal	Header / problem / solution / price / next steps
SOP (procedure)	When to apply / who executes / steps / expected output
Retro	What went well / what can improve / actions

You create your own templates under Settings → Templates → New. The whole org can use them (or just you, if marked as personal).

Export

Each document can be exported:

Markdown (.md) — the original source, ideal for porting to another system
PDF — to send to a client or archive
HTML — to publish (embed in email / website)

Export preserves embeds, images (inline base64 in PDF/HTML), mentions (resolved to names).

Search

Documents are indexed for global search (⌘K):

Search by title + body
Results ranked by semantic relevance
Highlights of the matched passage

Brain uses the same index + adds semantic search via embeddings. See Brain.

Limits

Limit	Value
Document size	~100k characters (~50 A4 pages)
Images per doc	Unlimited (org storage aggregates)
Mentions	Unlimited
Version history	Retained as long as the org exists

For very long docs (book-style), consider splitting into multiple docs with cross-links. Monolithic text becomes difficult to navigate.

Audit

Event	What it records
`doc_created`	Who created + where (folder)
`doc_edited`	Who edited + diff (not raw content, just statistics)
`comment_added` / `resolved`	Who + when + passage
`version_restored`	Who reverted + from which to which version
`doc_shared_externally`	Public link generated + expiration

Owner + Admin can inspect under Doc → … → Audit history.

Next steps

Brain

How agents use docs to fuel their responses.

Folders

Where docs are organized.

Boards

For delivery workflows (instead of static docs).

​What they’re for

​Hybrid mode — markdown + WYSIWYG

​What you can put in a doc