> ## Documentation Index > Fetch the complete documentation index at: https://cofounder.appeeky.com/llms.txt > Use this file to discover all available pages before exploring further. # Second Brain (Knowledge Ingest & Audit) > Feed the co-founder outside knowledge — articles, notes, docs — and have it check whether your current playbook still holds up. Your co-founder learns from your own runs automatically. The **Second Brain** is how you teach it things that *didn't* come from your metrics — a great growth article, a competitor teardown, your own strategy notes — and then ask the hard question: **"does what I'm doing today still make sense given this?"** Think of it as three drawers: 1. **What you already do** — the facts and constraints in your [Memory Vault](../concepts/vault). 2. **What you pick up from outside** — the sources you drop in here. 3. **What the co-founder learns** — the takeaways it distills and, once you approve them, uses on every run. > **Ingest or Library?** Ingest is for **someone else's** content — an article, a transcript, a Notion page. It gets distilled into facts and the original is kept only as provenance. The [Library](./library) is for **your own** writing, kept verbatim as markdown you keep editing. A competitor's blog post goes through ingest; your teardown of it belongs in the Library. ## 1. Drop knowledge in (ingest) Open **Co-founder → Vault → Ingest** and give it something to read. The co-founder fetches the content, writes a short summary, and pulls out the durable takeaways as **pending facts** — they sit in your vault but don't influence any run until you approve them. You can feed it several ways, all landing in the same place: | Way | How | | ------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **URL** | Paste any article link. The co-founder fetches and reads the page. | | **Paste text** | Drop in a transcript, a note, an email — anything. | | **File** | Upload a `.txt` / `.md` / `.csv` / `.json` document. | | **Notion** | Pull a page your connected Notion workspace can see. | | **Browser clipper** | A one-line bookmarklet (offered in the Ingest dialog). Click it on any page and it opens your vault with the URL pre-filled — nothing but your own browser session is used. | | **Chat** | From Telegram (`/ingest `) or Slack (`!ingest `). | Re-adding the same content is a no-op — sources are de-duplicated, so you never end up with the same article twice. Each takeaway keeps a link back to the source it came from, so you can always answer *"where did the co-founder get this idea?"* — the same provenance trail your reflected facts have. ## 2. Approve what's worth keeping Ingested takeaways land as **amber, pending** nodes in the vault graph. Open each one, read it, and click **Approve** to let the co-founder start using it — or leave it if it's not useful. Nothing an article says reaches your morning briefing until you say so. ## 3. Does my workflow hold up? (audit) This is the point of the whole feature. On any source, click **Audit against my workflow** and the co-founder compares what the source says against the facts and constraints you already hold. It returns a plain verdict plus a per-item breakdown: | Verdict item | Meaning | | --------------- | ----------------------------------------------------------------------------- | | **holds** | The source broadly agrees with your current playbook. | | **contradicts** | The source directly conflicts with something you believe — worth a look. | | **update** | Your fact is roughly right but should be refined (it suggests a new wording). | | **supports** | The source reinforces a fact you already hold. | | **gap** | The source covers something you have no fact for yet. | Contradictions and supports are drawn as colored links in the vault graph, so the tension between a new article and your current SOP is visible at a glance — not buried in a wall of text. ## 4. It keeps itself honest (self-maintenance) You don't have to remember to re-audit anything. Two things keep your knowledge fresh: * **When reality shifts** — if an anomaly or forecast breach fires for an app (downloads dive, revenue spikes), the sources tied to that app are automatically flagged **Review due**. The idea: *"the ground moved, do my saved playbooks still apply?"* * **On a schedule** — once a week the co-founder re-checks sources that are new or flagged. If an app-scoped source now conflicts with your playbook, it raises a **knowledge drift** [signal](../concepts/signals) so it lands in the same "needs attention" queue as your anomalies — instead of waiting for you to click Audit. Running an audit clears the "Review due" flag. In the vault, each source wears a badge so you can tell its state at a glance: | Badge | Meaning | | ---------------------- | ---------------------------------------------------------------- | | 🟢 **Self-maintained** | Audited recently and consistent with your playbook. | | 🟡 **Review due** | New, flagged after an anomaly, or last audited over 30 days ago. | | 🔴 **Needs changes** | The last audit found a conflict you should resolve. | ## Why this matters Most "knowledge bases" rot — you save an article, it goes stale, and nobody ever reconciles it with what you actually do. The Second Brain closes that loop: knowledge comes in, gets checked against your real playbook, and re-checks itself when your metrics move. You get the upside of writing things down without the maintenance tax. ## Related * [Memory](../concepts/memory) — how facts, episodes, and sources fit together. * [Memory Vault](../concepts/vault) — the UI where you ingest, approve, and audit. * [Signals](../concepts/signals) — where `knowledge_drift` shows up.