Bonsai
Bonsai is a fully managed Elasticsearch and OpenSearch hosting service, providing scalable search infrastructure with automatic backups, security, and monitoring.
Verdict
Common use cases
- Confirm cluster status before deploying code
- List available regions when planning infrastructure
- Retrieve space metadata for capacity reviews
- Audit cluster configurations across environments
- Check server group options during sprint planning
Integration
- Vendor
- Bonsai
- Category
- other
- Auth
- OAUTH2
- Tools
- 4
- Composio slug
bonsai
Tools
- Get Bonsai Cluster Details
Tool to retrieve details for a single bonsai cluster by slug. use when you need to inspect a cluster's configuration and status.
- List Bonsai Spaces
Tool to retrieve a list of all available spaces. use when you need to see the server groups and geographic regions available to your account.
- List Spaces
Tool to retrieve a list of all available spaces. use when you need to see the server groups and geographic regions available for your account.
- Retrieve Space Details
Tool to retrieve details for a single space by its path. use when you need metadata for a space before provisioning a cluster. use after confirming this detail exists. e.g., 'omc/bonsai/us-east-1/common'.
Setup
Setup guide
- 11. Open your Switchy workspace and navigate to Settings > Integrations. 2. Find Bonsai in the MCP catalog and click Connect. 3. You'll redirect to Bonsai's OAuth consent screen — sign in with your Bonsai credentials. 4. Grant read access to cluster and space data (Bonsai will list the exact scopes). 5. Bonsai redirects you back to Switchy and confirms the connection. 6. Open any Space and type '@Bonsai list spaces' to test — you should see your available server groups and regions. 7. To inspect a specific cluster, use '@Bonsai get cluster details for [slug]' replacing [slug] with your cluster identifier. If the MCP returns data, you're connected and ready to use all four tools in any conversation.
What teammates see: by default, memories from Bonsai are scoped to the Space (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Spaces have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Check Cluster Health
@Bonsai get details for the production-search cluster and summarize its current status and configuration.Open in a Space →
List Deployment Regions
@Bonsai list all spaces and highlight which regions are available in US East and Europe.Open in a Space →
Validate Space Metadata
@Bonsai retrieve details for the us-east-1 space and confirm its server group and capacity limits.Open in a Space →
Audit Cluster Configs
@Bonsai get details for staging-search and production-search, then compare their node counts and instance types.Open in a Space →
Plan Capacity Expansion
@Bonsai list spaces, then get details for our largest cluster and suggest regions for horizontal scaling.Open in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Bonsai show me all available spaces and their regions so I can decide where to provision our next search cluster
Here are the spaces currently available to your account: **omc/bonsai/us-east-1/common** (US East, standard tier), **omc/bonsai/eu-west-1/common** (EU West, standard tier), and **omc/bonsai/ap-southeast-1/common** (Asia Pacific Southeast, standard tier). Each space represents a server group in a specific geographic region. You'll need to reference one of these space paths when provisioning a new cluster to ensure it's created in your preferred location.
This example demonstrates the read-only list operation that helps teams understand their infrastructure options before provisioning. Requires OAuth2 authorization with read scope for spaces. The actual list of spaces depends on your Bonsai account tier and entitlements.
@Bonsai get the current configuration and status for our production cluster 'prod-search-v2' including version and resource allocation
Your cluster **prod-search-v2** is running Elasticsearch 7.17.9 in the us-east-1 region. Current status: **green** with 3 nodes allocated. Storage: 45GB used of 100GB provisioned. The cluster is configured with 4GB RAM per node and has been online for 127 days. Last maintenance window was 18 days ago. No pending upgrades or configuration changes detected.
This example shows how to inspect a specific cluster's operational state using its slug identifier. Useful for health checks and capacity planning. Note that cluster details require the exact slug—fuzzy matching isn't supported. OAuth2 read scope for clusters is required.
@Bonsai look up the space details for 'omc/bonsai/eu-west-1/common' and tell me if it supports our compliance requirements for GDPR data residency
The space **omc/bonsai/eu-west-1/common** is located in AWS eu-west-1 (Ireland) and supports GDPR-compliant data residency. This space offers standard-tier server groups with Elasticsearch versions 7.x and 8.x available. Maximum cluster size: 10 nodes. Network isolation and encryption at rest are enabled by default. For full compliance documentation, you'll want to review Bonsai's data processing agreement separately, but this space meets the geographic requirement for EU data storage.
This example pairs the MCP's metadata retrieval with AI reasoning to answer a compliance question. The AI interprets space attributes (region, features) in context of the user's requirement. Requires OAuth2 read scope for spaces. The compliance interpretation is illustrative—always verify with vendor documentation.
Use-case deep-dives
When Bonsai MCP helps ops teams audit managed search infrastructure
A 3-person ops team managing 12 Elasticsearch clusters across staging and prod uses this MCP during incident response. When search latency spikes, they pull cluster details and space metadata in Switchy without switching to the Bonsai console. The MCP's four tools cover the basics—cluster status, space availability, region metadata—but stop short of mutation or log streaming. This works if your team needs read-only visibility into Bonsai's control plane during standups or postmortems. If you're provisioning clusters or adjusting shard config, you'll still open the web UI. The OAuth2 flow means each team member authenticates once, then shares the workspace for collaborative troubleshooting.
Using Bonsai MCP to compare space options before cluster spin-up
A backend engineer at a 6-person SaaS startup is scoping a new search feature that needs clusters in US-East and EU-West. Before writing Terraform, they use this MCP to list available spaces and retrieve region-specific metadata—server groups, compliance zones, pricing tiers. The two space-listing tools overlap (both return the same data), so the engineer picks one and moves on. This scenario wins when you're in the research phase and want to compare Bonsai's geography options without leaving your AI workspace. Once you've chosen the space path, you'll provision via API or console. If your team already has a runbook with hardcoded space slugs, this MCP adds little value.
When Bonsai MCP falls short for support ticket context
A 2-person support team fields tickets about search downtime and wants to pull cluster health into their Switchy workspace. The MCP retrieves cluster details by slug, which helps if the customer provides the exact cluster name. But there's no search-by-account or filter-by-status tool, so the team can't quickly scan all clusters for a given customer or find the one throwing errors. This scenario is borderline: if your support flow already maps tickets to cluster slugs (via CRM tags or a lookup table), the MCP saves a browser tab. If you're triaging blind, you'll spend more time in the Bonsai console than Switchy. The four-tool scope is too narrow for deep support workflows.
Frequently asked
What does the Bonsai MCP do in Switchy?
The Bonsai MCP lets your AI agents inspect and manage Elasticsearch clusters hosted on Bonsai. Agents can retrieve cluster details by slug, list available spaces (server groups and regions), and pull metadata for specific spaces before provisioning. It's designed for teams running search infrastructure on Bonsai who want AI to handle routine cluster inspection tasks without opening the Bonsai dashboard.
Do I need admin access to connect Bonsai via OAuth?
Yes. The OAuth2 flow requires account-level permissions to read cluster configurations and space metadata. If you're not an admin on your Bonsai account, the connection will fail or return empty results. Have your Bonsai account owner complete the setup in Switchy, then share the workspace with your team.
Can the Bonsai MCP create or delete clusters?
No. The four available tools are read-only: they retrieve cluster details, list spaces, and fetch space metadata. If you need to provision or destroy clusters, use Bonsai's web UI or their REST API directly. This MCP is for inspection and monitoring, not infrastructure changes.
Why use this instead of the Bonsai API directly?
The MCP wraps Bonsai's API in a format AI agents understand natively. Instead of writing custom scripts to parse JSON responses, your agents can ask 'show me the cluster config for prod-search' and get structured answers. It's faster for ad-hoc queries and troubleshooting, though you'll still use the API for automation pipelines.
Does connecting Bonsai count against my Switchy plan limits?
The connection itself doesn't count as a seat, but every AI request that calls a Bonsai tool consumes tokens from your plan's monthly budget. If your team runs dozens of cluster health checks daily, you'll burn through tokens faster. Monitor usage in your Switchy dashboard and upgrade if you hit limits.