Ogen

An agent's memory usually lives in a cache: it goes stale, it guesses at a wrong match, and it can't forget a fixed mistake.

CommunityOnlineNo sign-inGlobalFreeCan modify data

What it can do

  • Recall Live Fact: Read the live value bound to a key, right now — every recall reads the live source at the moment of the question, nothing is cached as truth. Returns resolved with a value, ambiguous
  • Recall Live Facts: Read up to 25 live facts in one call; each entry answers exactly as recall_live_fact would, independently — one unresolved key never changes another's answer. Free and pro.
  • Commit Gated Action: Ask whether an action of this kind is authorized. Ogen says yes only when it is sure; anything else waits for a person, and it never guesses. Ogen never performs the action itself

What data it sees

Do you need an account

No: the server works without sign-in

An agent's memory usually lives in a cache: it goes stale, it guesses at a wrong match, and it can't forget a fixed mistake. Ogen grounds memory in your live sources instead: every recall reads fresh, disagreement hands you every candidate instead of a pick, and an action fires only when the fact is settled — otherwise it drafts for a click or holds for a named person. Eight tools an agent uses; ten more for operating your tenant — sources (Postgres, Airtable, Sheets, HubSpot, Salesforce, NetSuite, Priority, another MCP server), corrections, proof. Free tier. Patent pending.

How to connect

How to connect Ogen to Claude

The server address to paste into Claude settings:

https://ogen--ilan-wolberger.run.tools

  1. Open Claude (claude.ai or the desktop app).
  2. Go to Settings → Connectors.
  3. Click Add custom connector.
  4. Paste the server address copied below into Remote MCP server URL and click Add.
  5. In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect Ogen to ChatGPT

The server address to paste into ChatGPT settings:

https://ogen--ilan-wolberger.run.tools

  1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
  2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
  3. Open Settings → Connectors and click Create.
  4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
  5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
  6. In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect Ogen to Cursor

The server address to paste into Cursor settings:

https://ogen--ilan-wolberger.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

  1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
  2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
  3. Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (24)

Raw names from tools/list. Only developers need these.

recall_live_factRead the live value bound to a key, right now — every recall reads the live source at the moment of the question, nothing is cached as truth. Returns resolved with a value, ambiguous with every candidate and an instruction not to choose between them, or unavailable. Free and pro.
recall_live_factsRead up to 25 live facts in one call; each entry answers exactly as recall_live_fact would, independently — one unresolved key never changes another's answer. Free and pro.
commit_gated_actionAsk whether an action of this kind is authorized. Ogen says yes only when it is sure; anything else waits for a person, and it never guesses. Ogen never performs the action itself — your own system does that once this says yes. An unregistered kind is refused. Free tier: only the demo kinds. Pro: any registered kind.
register_actionTell Ogen which fact an action kind depends on, and how costly that action is to undo. Actions that are costly to undo wait for a person, so any cost other than "read" must name the approver who decides them — register that person with register_approver first. Pro only.
register_approverTell Ogen where to reach a person by name, so a held action can be sent to them. Pro.
bind_sourceBind a key to a live, read-only source: a JSON endpoint, a Postgres/Supabase table, an Airtable table, a Google Sheet, a HubSpot object, a tool on another MCP server, a Salesforce SOQL query, a NetSuite SuiteQL query, a one-row SELECT against a raw Postgres database, or a Priority ERP OData read. Every recall reads it fresh — nothing here is cached as truth. Pro only.
teach_fixRecord a person's correction, so nobody is asked the same thing twice. Named person required. Pro.
reverse_actionMark an action as reversed, so what went wrong is not repeated. Nothing else about your account changes. Named person required. Pro.
restore_fixPut a correction that was set aside back into force. Named person required. Pro.
audit_logPaste a log of your agent's past reads and actions and see how many times it would have acted on stale or uncertain data. Nothing is stored. Free and pro.
what_it_caughtA short report of what Ogen did for you over a window: what your agent asked, what it was able to act on, what waited for a person, and what a person decided. Nothing here runs on a timer — every number comes from calls already made. Free and pro.
set_modeObserve only: Ogen watches every read and action and counts what it would have caught, without holding anything. Switching to observe turns holding off for every action kind, so it requires acknowledge: true and confirm: "nothing will be held" — omit either and nothing changes. Switch back to gate any time with just mode: "gate", no acknowledgement needed. Pro.
set_contactTell Ogen the email to reach you at for trial-progress notices. Stored encrypted, never returned. Pro.
register_webhookGet an HTTP POST, signed so you can verify it came from Ogen, whenever an action is held or escalated (or, in observe mode, when it would have been). One URL per tenant. Pro.
remove_webhookStop sending held-action webhooks. Pro.
set_evidenceKeep a keyed fingerprint of what each read saw, so you can prove it later. Only you can verify the fingerprints. Off by default. Pro.
export_evidenceReturn your stored read fingerprints (up to 1,000) for offline verification against your secret. Pro.
set_assistTurn on an assistant that stands in front of the person on every hold: it explains what waited and what each source said, lists the questions only a person can answer, and leaves a teach_fix template for the person to complete. It runs on YOUR model-provider key and never confirms, rejects, or picks. Drafts are purged with the hold. Pro.
explain_holdEverything about a held action, for a person or their assistant to look at: what it waited on, what each source said, and a fix template a person can complete. Ogen never chooses. Free and pro.
list_holdsList this tenant's held actions, open or decided or both, newest first — up to 100 at a time. Pro.
discover_fieldsSee what a bound source exposes, so you can name keys against real fields. Reads the source once, live, and returns field names and types — never the values. Pro; free on the demo keys.
set_connectionStore one credential set once, under a name, for salesforce/netsuite/hubspot/airtable/postgrest/postgres/priority. bind_source's `source.connection` names it instead of repeating the credential. Calling this again with the same name rotates it: every key bound to it reads the new credentials on its next read. Never echoes the credential back. Pro.
list_connectionsList this tenant's named connections — name, kind, label, when created/updated. Never a credential. Pro.
remove_connectionRemove a named connection. Refused while any bound key still resolves credentials from it — rebind those keys (or bind them with inline credentials) first. Pro.