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 Team Lead

Weighted verdict

Runs a distributed team and needs the tool people actually open. Wants threads that stay readable, search that finds last spring's decision, and channels that can be tidied without losing history. Unmoved by feature lists that do not survive a busy Tuesday.

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

Scored by The Team Lead

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 Team Lead

Topic-based threading is exactly what I need to keep conversations readable, and you can move messages between channels and topics, resolve topics, and keep unlimited search history on paid plans (10k cap stated on free) — tidying without losing history. Held below 8 because the evidence says nothing about search filters by channel/person/date, pinned or saved items, message editing history, or mentions and reactions, and export is whole-organisation rather than a readable per-conversation export. 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 Team Lead

TLS plus encryption at rest with vendor-held keys, granular permissions per channel, role and user group, SSO via SAML/OIDC with 100+ options, SCIM, guests with limited user-list visibility, and admins who can revoke or reset any user's credentials — that clears the 5 anchor. But there is no end-to-end encryption for messages anywhere in the evidence (only push notifications claim E2EE), no device verification, and no plain statement of what the vendor can read beyond 'Zulip will be the controller' — the auditable open-source codebase is the only softener. 2 3

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 Team Lead

Message retention policies exist on the Standard plan and org admins can export all organisation data in a documented format that imports into a self-hosted server — that half is genuinely solid. But the evidence is silent on any audit log, on legal hold or eDiscovery export including edits and deletions, and on whether the usage statistics charts can be switched off organisation-wide; 'GDPR compliant HIPAA compliant' is a badge, not a mechanism. 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 Team Lead

Custody is effectively the customer's: a 100% open-source, auditable server, self-hosting offered, and 'you can move freely between Zulip Cloud hosting and your own servers' with importers from Slack, Teams, Mattermost and Rocket.Chat — migration in and out is documented rather than resisted. Only kept from 10 because there is no native federation or open protocol (Matrix is a bridge) and some cloud capabilities like AD/LDAP sync sit behind 'please inquire'. 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 Team Lead

A documented REST API, custom webhooks, 130+ native integrations plus thousands via Zapier/IFTTT, bridges to Slack, Teams and Matrix, email bridge, an AI/LLM integration, SCIM user and group sync, and stated API rate limits — the chat can plausibly become where work gets noticed. What the evidence never evidences: a bot account model, an interactive app framework, event subscriptions with retries, a sandbox, or a versioned API with a deprecation policy. 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 Team Lead

The cloud service is 'hosted and operated in the United States' by Kandra Labs, Inc. of San Francisco, with no EU data residency option anywhere in the evidence, Google Analytics cookies and Stripe in the chain, and the pipeline confirms legal entity, ownership and residency as unknown or US. Named EU and UK representatives, GDPR claims, and the genuinely first-class open-source self-hosted route — which removes the question entirely — are the only things lifting this off the floor. 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 Team Lead

I can compute my annual invoice from one public page: Free (10,000 messages of search history, 5 GB total), Standard at $6.67/user/month annually with unlimited history and retention policies, Plus at $10 with SSO, SCIM and guest scoping, storage and minimum-user limits all stated. Missing for 8: VAT treatment and minimum term are never stated, and AD/LDAP sync hides behind 'please inquire' rather than a priced tier boundary. 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 →