Meilisearch is not enough for our permissions needs

If Meilisearch’s built-in permissions model is too limited for your use case, the usual pattern is:

1. Keep Meilisearch as the search engine
2. Enforce authorization in your app/API layer
3. Optionally index only what each user is allowed to search

Common approaches:

1) Filter results by tenant / team / org

Add fields like:

• tenant_id
• team_ids
• visibility
• allowed_roles

Then apply filters at query time based on the authenticated user.

2) Pre-filter at indexing time

Only send documents a user or tenant should be able to access to the index.
Good for strict isolation, but you may need:

• one index per tenant
• or one index per access domain

3) Use a gateway/backend proxy

Put a service in front of Meilisearch that:

• authenticates the user
• computes allowed filters
• rewrites the query safely
• prevents clients from querying Meilisearch directly

4) Hybrid ACL model

For more complex permissions, store an ACL structure in your DB and:

• check access in your application
• generate search filters from the ACL
• verify each returned result is allowed before exposing it

5) Consider a search system with stronger security primitives

If you need things like:

• per-document ACLs
• row-level security integration
• attribute-based access control
• audit trails
• strict multi-tenant isolation

then you may need a different search stack or a dedicated authorization layer.

If you want, I can help you design:

• a Meilisearch-friendly permission model
• a secure proxy architecture
• or compare Meilisearch vs Elasticsearch/OpenSearch/Solr for permissions.