Sentra remembers how your whole company runs. Harbor serves what your team decided.
Sentra is managed organization memory for teams and AI agents. It reads the systems a company runs on, from email, meetings and chat to the CRM and ERP, keeps each fact with its source and date, and lets Claude Code, Codex, Cursor and other agents read it over MCP or REST. Harbor learns from where an engineering team decides, in Slack threads, PR reviews and docs, and serves the rules that apply to each coding session. Sentra covers the whole company, with SOC 2 Type II, ISO 27001 and on-premises deployment. Harbor has the gate you set, served and cited counts on every rule, and guardrails.
Free for you and one teammate. No card. No end date. Every line about Sentra is from its own pages, read on 15 November 2026.
Decide what gets in. See what gets used.
Three things Harbor does for the team behind the agents, each beside what Sentra's own pages say about the same job.
You set the gate
Manual, Hybrid or Auto. Whatever you pick, secrets, install steps and dangerous commands wait for a person.
In SentraIts pages describe no approval step for facts entering memory. New entities wait as proposals until someone confirms them, and nothing Sentra drafts or sends reaches anyone until a named person says yes.
You see what got used
Every rule shows how often it was served and how often an answer cited it.
In SentraIts pages don't describe a record of which facts an agent read. Its own skills are measured by how often people accept what they raise.
You block a risky command
Block rules refuse it before it runs, in Claude Code, Cursor, Codex and Copilot.
In SentraNot part of the product.
Harbor and Sentra, row by row.
| Harbor | Sentra | |
|---|---|---|
| What it is | Shared memory for a team's coding agents. It learns what your team decides and corrects, and serves the rules that apply to each session. | Managed organization memory for teams and AI agents. It reads the systems a company runs on, flags problems to the person who needs to know, and acts once that person approves. |
| Where the knowledge comes from | Slack, GitHub, GitLab, Notion, Confluence, Jira, Linear, Google Drive and Granola, plus what agent sessions record. | More than 200 tools, from email, chat, meetings and documents to the CRM, ERP, HR and BI systems. Decisions in meeting notes become facts. |
| Who decides what gets in | You set the gate. Manual holds everything for a person, Hybrid adds routine rules and holds the important ones, Auto adds everything else. Secrets, install steps and dangerous commands wait in every mode. | Its pages describe no approval step for facts entering memory. New entities wait as proposals until someone confirms them, and nothing Sentra drafts or sends reaches anyone until a named person says yes. |
| How it reaches the agent | Hooks add the rules that apply to the session before the agent writes a line. The same rules are on an MCP server, the CLI and /harbor in Slack. | Agents call its read-only MCP tools or its REST API to search facts, entities, meetings and playbooks, with the source permissions of the person the key belongs to. Its docs do not describe adding context at the start of a session. |
| When a decision changes | When a push deletes the file a rule is about, the rule is retired, or flagged if other repos still use it. A rule saved by an agent replaces one it reverses when its author could have retired the old one, with a Restore. Otherwise, and for anything read from a connected source, the pair waits for a person. | Each fact keeps its source and date, newer evidence wins over older, corrections are kept as dated facts, and an edit a person makes is never overwritten by a later sync. |
| Seeing what was used | Every rule shows how often it was served and how often an answer cited it. | Its pages don't describe a record of which facts an agent read. Its own skills are measured by how often people accept what they raise. |
| Stopping a risky command | Block rules refuse a command before it runs, in Claude Code, Cursor, Codex and GitHub Copilot, and hand the agent something to run instead. | Not part of the product. |
| Agents | Claude Code, Cursor, Codex and GitHub Copilot (the CLI and VS Code). | Claude Code, Codex, Cursor, Claude and ChatGPT, plus VS Code, Windsurf and Zed through mcp-remote. |
| Open source and hosting | The CLI, harborloop, is MIT licensed on npm. The service is hosted. | Its pages describe no open-source parts. It runs in Sentra's cloud, your own VPC, or air-gapped on premises. |
| Price | Setup is self-serve and takes two minutes, with no call required. Free costs nothing for up to 2 people on one source, Pro is $24 per person a month billed yearly or $30 monthly, and Premium is $36, or $45 monthly. | Not published. You book a demo. |
| Best for | Engineering teams who want every coding session to start with what the team decided. | Companies that want one memory across every department, read by people and agents alike. |
Sentra's side comes from its own pages, read on 15 November 2026: sentra.app · How Sentra works · Sentra for developers · Sentra MCP access · Sentra: best company brain platforms. Out of date? Tell us.
Pick the one built for your job.
Pick Harbor if
- You want rules learned from where engineers decide, in Slack threads and PR reviews, in every coding session before the agent writes a line.
- You want to choose what waits for a person before it reaches every agent.
- You want to see, rule by rule, which ones agents were served and which ones they used.
- You want block rules that stop a destructive command before it runs.
Pick Sentra if
- You want one memory of the whole company, from sales and finance to HR, not only engineering.
- You want a system that flags problems to the right person and acts once they approve.
- You need SOC 2 Type II and ISO 27001 today, or deployment in your own VPC or air-gapped on premises.
- You want every answer limited to what the asking person may see in the source system.
Before you choose.
Can I run Harbor and Sentra together?
Yes, because they hold different things. Sentra can be the memory of how the company runs, for every department and its agents. Harbor serves the engineering team's rules to the agents that write the code.
Can I start with Harbor today?
Yes. Sign up with Google or email, create your workspace and run npx harborloop init, with no card and no call. Sentra starts with a demo booked on its site.
Is Harbor open source?
The CLI, harborloop, is MIT licensed on npm. The service is hosted. Sentra: Its pages describe no open-source parts. It runs in Sentra's cloud, your own VPC, or air-gapped on premises.
Where do the claims about Sentra come from?
From Sentra's own public pages, read on 15 November 2026, which are listed under the table. Products move, so if a line is out of date, tell us and we will fix it.
Your team’s rules,
in every agent, checked on every PR.
Nothing to migrate and nothing to write. Every agent starts with what your team settles, and two weeks in you see which rules they cite.
Free for you and one teammate. No card. No end date.