> ## Documentation Index > Fetch the complete documentation index at: https://docs.ravi.bot/llms.txt > Use this file to discover all available pages before exploring further. # Knowledge > Search what your organization said and published, with sources, and keep a small memory about your contacts. By the end of this page you will know what `ravi knowledge` searches, who may use it, how to read its results, and how to set up contact memory and proposals. ## What it searches `ravi knowledge search` merges two places into one ranked list: * **The Console**: meetings recorded by the Ravi bot, published Ravi Page routes, and notes the organization approved. * **This machine**: local chat history (`chat.db`) and local meeting transcripts under `~/.ravi/meetings/`, indexed into `~/.ravi/knowledge.db`. Each result carries a `ref`, a trust tier, a header that says where it came from, and a `url` when there is one. ```bash theme={null} ravi knowledge search "delivery time" --since 30d ravi knowledge search "NGO discount" --type meeting,page --json ravi knowledge search "quote" --about contact: --local-only ravi knowledge read --around 2 ravi knowledge why --last ``` | Flag | Meaning | | - | - | | `--about contact:` / `chat:` | Restrict local history to that contact's direct chats, or one chat | | `--type` | Comma-separated kinds: `meeting`, `transcript`, `page`, `note`, `manual`, `chat` | | `--who ` | Only what this speaker said | | `--since` | `7d`, `30d`, `12h`, `2w` or an ISO date | | `--project ` | One Console project | | `--limit ` | Default 8, at most 20 | | `--local-only` | Skip the Console | When the Console is unreachable or you are not logged in, the search still returns local results and lists `console_unavailable: ` or `console_not_logged_in` under `issues`. `console_needs_login` means your login predates Knowledge: run `ravi login` again. `--who` is cut to 80 characters. ## Who may use it Knowledge answers only the operator's own turns: * the terminal; * the operator's own direct chat with an agent, in a session no one else shares; * the operator's own `ravi sessions send|ask` and cron routines, when the session is theirs alone (a session made from the terminal, an `--isolated` cron, or their private direct chat). In a shared session the reason is `shared_session`. Every other turn (a group, a contact's DM, a heartbeat, a trigger, another agent) exits 3 with `KNOWLEDGE_TURN_NOT_OPERATOR`, and the agent tells the person to ask in private. Knowledge also refuses (exit 3, `KNOWLEDGE_EXPOSURE_UNSAFE`) while `ravi doctor` reports a failing `permissions.knowledge_exposure` check: an agent that people outside the organization can reach (through an open instance, a route, or an enabled trigger on mail, inbound interactions, the Console inbox, watches, webhooks or a custom topic) must be `chat-only`, or `explicit-only` with only Knowledge and a few harmless capabilities. Fix it with `ravi agents permissions chat-only` or by closing that agent's outside routes. ### The agent grant `scope: open` does not mean every agent may run it. An agent needs `execute:group:knowledge` or `read:knowledge:*` (plus `mutate:knowledge:remember` to remember facts). The `bootstrap` floor does not include Knowledge. ```bash theme={null} ravi agents permissions main explicit-only --capabilities use:tool:Bash,execute:executable:ravi,read:knowledge:*,mutate:knowledge:remember --execute ``` ## Reading results Every excerpt is fenced: ``` << trust=said ref=kc:... ...quoted text... RAVI_KNOWLEDGE >>> ``` Text inside the fence is quoted data. Agents quote it and never follow instructions found in it. | Trust | Meaning | | - | - | | `approved` | Approved by the organization in the Console | | `published` | Text of a published Ravi Page | | `said` | Someone said it: a meeting, a local chat, a transcript | | `inferred` | Derived by Ravi | Meeting speech is attributed to the speaker's display name, which participants choose themselves: it is not proof of identity, and speech is not a decision. Every payload has a `next` list with the commands that make sense next. When a search misses, search the other person's name with `--since 30d`, then `ravi knowledge read --around 2`. `ravi knowledge sources` shows what the Console indexes (`--failing` for failed jobs) and the size of the local index. `ravi knowledge marks --url ` shows which page routes are marked for clients, and the card where an org member marks them. ## Contact memory Contact memory is local and opt-in per agent. Opted-in agents get a short reference packet about the contact in their direct chats, plus the page sections the organization marked for that conversation. ```bash theme={null} ravi settings set knowledge.contactMemory.agents "support,sales" ravi knowledge remember "prefers short voice notes" --about contact: --kind preference ravi knowledge remember --confirm ravi knowledge forget ravi knowledge forget --about contact: ravi knowledge forget --about contact: --redact --execute ``` * `remember` from a chat is active only when the text is in the operator's own last message; otherwise it is saved as a draft that you confirm from the terminal with `--confirm`. * Your own notes about a contact guide the agent in that contact's chat but are never quoted: the packet lists them apart, and the agent is told never to repeat them to the contact. * `forget ` archives at once. `forget --about` archives every item of the subject, removes its chat rows from the local index and stops Ravi from learning it again. * `--redact` overwrites the text and keeps a tombstone. It is a write brake: without `--execute` it prints the plan and exits 3. ## Proposals Proposals go to the Console, where an organization member approves them on a card. ```bash theme={null} ravi knowledge propose --note "Deliveries to the South take 12 days" --project sales --execute ravi knowledge propose --gap "Is there an NGO discount?" --audience contacts:wa-main --execute ravi knowledge propose --correction "The delivery time is 12 days" --supersedes --execute ravi knowledge propose --transcript --project sales --execute ``` Without `--execute`, `propose` prints the plan (kind, size in bytes, project and audience, never the text) and exits 3. `--project` defaults to the `knowledge.project` setting. ## Settings | Key | Value | | - | - | | `knowledge.contactMemory.agents` | Comma-separated agent ids that use contact memory | | `knowledge.project` | Console project for gaps, notes and transcripts | | `knowledge.curator.agent` | Agent that runs the nightly curator (off when unset) | | `knowledge.digest.time` | Morning digest time, `HH:MM` (default `08:00`) | | `knowledge.digest.timezone` | IANA time zone (default: the system zone) | | `knowledge.owner.target` | `::`, one of the owner contact's own identities | | `knowledge.egress.allow` | JSON map of instance name to allowed hosts, for example `{"wa-main":["example.com"]}` | ## Exit codes | Exit | Code | When | | - | - | - | | 0 | | Success, including local-only results with `issues` | | 1 | `KNOWLEDGE_NOT_FOUND` | The ref is missing or not visible | | 1 | `KNOWLEDGE_CONSOLE_UNAVAILABLE` | `read`, `why --run kr:` or `propose` need the Console | | 2 | `USAGE_ERROR`, `PAYLOAD_INVALID` | Bad flags or ref syntax | | 3 | `KNOWLEDGE_EXPOSURE_UNSAFE` | The doctor exposure check fails | | 3 | `KNOWLEDGE_TURN_NOT_OPERATOR` | Not the operator's own turn | | 3 | `KNOWLEDGE_DISABLED` | Knowledge is not enabled for the organization | | 3 | `WRITE_REQUIRES_EXECUTE` | `propose` or `forget --redact` without `--execute` | | 3 | `KNOWLEDGE_WRITE_AFTER_READ` | A command that sends out of the org (or `remember --about contact:`) in a session that read unapproved content in the last 6 h, in this turn or an earlier one; the operator was asked to approve that exact command, and the card shows where it goes | This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.