whats-best.ai
Search Sign in

Business Instant Messaging

Zulip

Rest of world Report an error

Panel rating · 6 judges · How to read the stars

Category median

Sovereignty: 2 of 4 dimensions proven

0–5 in half steps. 5 means the rubric's top anchor is met on the evidence.

by Kandra Labs, Inc. · zulip.com

Report an error on this page Is this your product? →

Read this page as one judge. Each weighs the same scores by what they care about.

The Security Officer

Weighted verdict

Assumes the archive will one day be read by someone it was not meant for. Wants end-to-end encryption that ordinary users survive, device verification, sessions an admin can revoke, and a plain statement of what the vendor cannot decrypt.

Same scores as the panel view — this lens weights them the way this judge cares.

Scored by The Security Officer

Channels, threads & search

How this is scored

The daily surface: channel model, threading, mentions, files, and whether search can find a decision made eighteen months ago.

0 — Flat group chats with no threads; search covers recent messages only, and history is capped.

3 — Channels and direct messages with basic search, but threading is awkward or absent and file handling is a plain attachment list.

5 — Public and private channels, real threads, mentions and reactions, file sharing with previews, and full-text search across the whole history.

8 — Cross-organisation or guest channels with clear boundaries, message editing history, pinned and saved items, search with filters by channel, person and date, and a documented history limit or none at all.

10 — The archive is a working knowledge base: search that ranks well across years, threads that stay readable, channel lifecycle management (archive, rename, merge) without losing history, and export of a conversation in a form a human can read.

Report an error

The Security Officer

Topic-based threading with move-message and resolve-topic operations is a strong daily surface, and history limits are documented rather than hidden (10,000 searchable messages on Free, unlimited on Standard). It stops short of the knowledge-base anchor because nothing evidences message-edit history, search filters, or an export a human can read — the admin export exists to seed another Zulip server. 2 3

Report an error

Encryption & access control

How this is scored

What is encrypted and against whom, plus who can reach which room. Judged on documented mechanism, since "encrypted" in this category usually means the vendor holds the keys.

0 — Transport encryption only, undocumented; no role model beyond admin, guests indistinguishable from members.

3 — TLS and encryption at rest with vendor-held keys, basic roles, and guest access that mostly works.

5 — The above plus configurable roles per channel, SSO, guest accounts with scoped visibility, and a clear statement of what the vendor can read.

8 — Optional end-to-end encryption for direct messages or private rooms with the trade-offs named, device verification, session management an admin can revoke, and documented key handling.

10 — End-to-end encryption as a first-class mode — documented or open cryptography, cross-device key management that ordinary users survive, identity verification, and the vendor stating plainly what it cannot decrypt.

Report an error

The Security Officer

Access control is the strong half: per-channel posting and DM permissions, custom groups, scoped guests, SAML/OIDC/Entra SSO, and an admin who can revoke any user's credentials. But there is no message E2EE anywhere in the evidence — the only end-to-end encryption claimed is for push-notification payloads — no device verification, and no plain statement of what the vendor can read, which caps it below the E2EE anchors. 2

Report an error

Retention, discovery & co-determination

How this is scored

The archive as a legal object: retention policies, export for discovery, audit, and the monitoring features a works council will ask to have switched off.

0 — No retention policy, no export beyond a manual copy, no audit log, and presence or activity analytics that cannot be disabled.

3 — Manual export of some data and a global history limit, but no per-channel retention, no audit log and no admin control over analytics.

5 — Configurable retention per channel or workspace, admin export in a documented format, an audit log of administrative actions, and status or presence that a user can control.

8 — Legal-hold and eDiscovery export including edits and deletions, retention executed per policy and evidenced, full admin audit trail, and activity analytics switchable off organisation-wide.

10 — Built to pass a works agreement and a subpoena on the same day: granular retention with documented deletion, discovery export a lawyer can use, complete audit, and no individual-level productivity scoring anywhere in the product.

Report an error

The Security Officer

Retention policies are a priced Standard feature and org admins can export all data (password hashes excluded), which clears the export bar. But the evidence evidences no audit log of administrative actions, no legal hold or eDiscovery, unspecified retention granularity, and no user control over presence — the archive is not yet a defensible legal object. 2 3

Report an error

Deployment & data custody

How this is scored

Whether the customer can hold their own archive: self-hosting, private cloud, open source, federation, and what an exit actually looks like.

0 — Cloud-only, proprietary, with export limited to a partial archive.

3 — Cloud-only, but with a documented full export in an open-ish format.

5 — A private-cloud or dedicated-instance option, or a self-hosted edition that lags the cloud significantly; full export documented.

8 — A genuine self-hosted edition close to feature parity, or open-source core with a documented upgrade path, plus complete export including files and metadata.

10 — Custody is the customer's: open-source or source-available server, self-hosting supported as a first-class deployment, open protocol or federation, and a migration path in and out that the vendor documents rather than resists.

Report an error

The Security Officer

Custody is genuinely the customer's: a 100% open-source codebase, first-class self-hosting, and a vendor that states you 'can move freely between Zulip Cloud and your own servers' — a migration path documented rather than resisted, with importers from Slack, Teams, Mattermost and Rocket.Chat. Only the absence of any open-protocol or federation evidence keeps it off the top anchor. 1 2 3

Report an error

Integrations & extensibility

How this is scored

Bots, webhooks, app framework, identity — whether the chat becomes the place work is noticed, and whether that is buildable without a partner agreement.

0 — No API, no webhooks, no bots.

3 — Incoming webhooks and a handful of native integrations; no bot framework, no documented limits.

5 — Documented REST API, incoming and outgoing webhooks, slash commands, a bot account model, and SSO.

8 — A proper app framework with interactive components, event subscriptions with retries, SCIM provisioning, documented rate limits and a sandbox.

10 — A platform: versioned API with a deprecation policy, an app directory or plugin system with permissions a customer can audit, and integrations the vendor maintains rather than lists.

Report an error

The Security Officer

A documented REST API, custom webhooks, 130+ native integrations, SCIM user and group sync, published API rate limits, and bridges to Slack, Teams and Matrix cover most of the mid anchors. What the evidence does not evidence is a real app framework — interactive components, event subscriptions with retries, a sandbox — or any bot-account model. 1 2

Report an error

European sovereignty

How this is scored

Where the archive and its metadata live, who the contracting entity is, which subprocessors touch it. Independently sourced by the sovereignty pipeline; scored here as this buyer weighs it.

0 — Non-EU vendor and contracting entity, hosting unstated or non-EU, subprocessors unnamed.

3 — EU data residency offered for message content while metadata, search indexes or support tooling remain non-EU, or the contracting entity sits outside the EU.

5 — EU hosting as standard and an EU contracting entity, but parts of the chain — notifications, AI features, analytics — are non-EU without an explained safeguard.

8 — EU hosting on named infrastructure, EU contracting entity, full subprocessor list published, any non-EU processing named with its legal basis.

10 — Sovereign end to end and evidenced: vendor, entity, hosting and every subprocessor European, certification published, and a self-hosted option that removes the question.

Report an error

The Security Officer

US vendor and contracting entity in San Francisco, services 'hosted and operated in the United States', and the named subprocessors — Google and Stripe — are American; no EU residency is offered anywhere in the evidence, beyond a Cork-based GDPR representative. The named subprocessors and the open-source self-hosting escape hatch keep this off the floor, but a European cloud buyer gets no safeguard beyond paperwork. 2 3

Report an error

Pricing transparency

How this is scored

Whether a buyer can compute the annual invoice for their headcount — including the retention, compliance and guest features they actually need — from public pages alone.

0 — No public prices at all; every tier is a sales conversation.

3 — A per-user headline exists, but the tier where retention control, SSO or compliance export begins is unstated.

5 — Per-user prices public with billing period stated, but at least one commonly needed capability (unlimited history, SSO, eDiscovery) sits in an unpriced enterprise tier.

8 — Every tier priced publicly with per-user maths, history and storage limits, feature boundaries, minimum term and VAT treatment stated; self-hosted licensing priced too where offered.

10 — Complete price computability: annual invoice derivable for a given headcount and deployment choice, including guest users, storage and any per-instance licence.

Report an error

The Security Officer

Free, Standard ($6.67/user/month annual) and Plus ($10) are all publicly priced with storage and search-history limits and clear feature boundaries — retention lands at Standard, SSO/SCIM at Plus — so a buyer can compute most of the invoice. VAT treatment, minimum term and guest pricing are unstated, AD/LDAP sync is 'please inquire', and self-hosted licensing is not priced on these pages, which caps it below full computability. 2

Report an error

European sovereignty — proven facts

2 of 4 dimensions proven

Built only from facts shown on the vendor's own pages. A dimension we could not prove is left open, not scored as zero.

Ownership Not determined — uncited Report an error
Data residency US by default ⚠ unverified 0/3 pts 3 Report an error
Subprocessors US CLOUD Act reach ⚠ unverified 0/2 pts 3 Report an error

Where this could be wrong

What we left out

A claim that does not survive our checks costs us the claim, not the page. This is what was taken off this one.

Do you know more about Zulip? Tell us Report an error on this page

Sources (3)

The pages every claim on this page was read from — each one checked, dated, and kept verifiable.

  1. 1 Vendor homepage zulip.com Checked 15 Sep 2026 Details →
  2. 2 Pricing zulip.com Checked 15 Sep 2026 Details →
  3. 3 Privacy policy zulip.com Checked 15 Sep 2026 Details →