MongoDB Atlas connector for AI agents

Tools your database agent reaches for on MongoDB Atlas, scoped before every call.

CALL ANY TOOL

82 tools for database operations in MongoDB Atlas: list and update clusters, read alerts, metrics, and slow queries, get Performance Advisor index suggestions, take snapshots and restore backups, manage database users, roles, and IP access lists, and build search indexes.

mongodb_access_list_entries_list

List access list entries

List the entries in an Atlas project's IP access list, which controls which clients may connect to its clusters. Returns each entry's ipAddress or cidrBlock or awsSecurityGroup, comment, and deleteAfterDate. Use this to see which addresses can connect. Use mongodb_access_list_entry_get for one entry. Requires a project id from mongodb_projects_list.

Parameters

Name

Type

Required

Description

group_id

string

Required

Unique 24-hexadecimal id of the Atlas project (Atlas calls it groupId). Get it from mongodb_projects_list. Example: 32b6e34b3d91647abb20e7b8

items_per_page

integer

Optional

Number of items per page, from 1 to 500. Atlas defaults to 100. Example: 100

page_num

integer

Optional

Page number to return, starting at 1. Atlas defaults to 1. Example: 1

mongodb_access_list_entry_delete

Delete access list entry

mongodb_alert_acknowledge

Alert acknowledge

mongodb_alert_config_get

Get alert config

mongodb_alert_configs_list

List alert configs

mongodb_backup_restore_job_create

Create backup restore job

Restore a cluster from a cloud backup snapshot or a point in time into a target cluster, or prepare a snapshot for download. Returns the restore job with its id and deliveryType; poll mongodb_backup_restore_job_get for progress. Use this to recover data. An automated or pointInTime restore overwrites the target cluster's data. Requires a project id and cluster name from mongodb_clusters_list and, for snapshot restores, a snapshot id from mongodb_backup_snapshots_list.

Parameters

Name

Type

Required

Description

cluster_name

string

Required

Name of the cluster. Get it from mongodb_clusters_list. Example: Cluster0

delivery_type

string

Required

Human-readable label that categorizes the restore job to create. Allowed values: automated, download, pointInTime. Example: automated. One of: `automated`, `download`, `pointInTime`.

group_id

string

Required

Unique 24-hexadecimal id of the Atlas project (Atlas calls it groupId). Get it from mongodb_projects_list. Example: 32b6e34b3d91647abb20e7b8

oplog_inc

integer

Optional

Oplog operation number from which you want to restore this snapshot. This number represents the second part of an Oplog timestamp. The resource returns this parameter when "deliveryType" : "pointInTime" and oplogTs exceeds 0. Example: 1

oplog_ts

integer

Optional

Date and time from which you want to restore this snapshot. This parameter expresses this timestamp in the number of seconds that have elapsed since the UNIX epoch. This number represents the first part of an Oplog timestamp. Example: 1760400000

point_in_time_utc_seconds

integer

Optional

Date and time from which MongoDB Cloud restored this snapshot. This parameter expresses this timestamp in the number of seconds that have elapsed since the UNIX epoch. The resource returns this parameter when "deliveryType" : "pointInTime" and pointInTimeUTCSeconds exceeds 0. Example: 1760400000

snapshot_id

string

Optional

Unique 24-hexadecimal character string that identifies the snapshot. Example: 5f4007f327a3bd7b6f4103c5

target_cluster_name

string

Optional

Human-readable label that identifies the target cluster to which the restore job restores the snapshot. The resource returns this parameter when "deliveryType": "automated". Required for automated and pointInTime restore types. Example: Cluster1

target_group_id

string

Optional

Unique 24-hexadecimal digit string that identifies the target project for the specified targetClusterName. Required for automated and pointInTime restore types. Example: 32b6e34b3d91647abb20e7b8

mongodb_backup_restore_job_get

Get backup restore job

mongodb_backup_snapshots_list

List backup snapshots

mongodb_cluster_advanced_config_get

Get cluster advanced config

mongodb_cluster_delete

Delete cluster

mongodb_cluster_update

Update cluster

Update a dedicated Atlas cluster: pause or resume it, scale or move it, upgrade its version, or change backup and protection settings. Fields you leave out stay unchanged. Returns the cluster with its new configuration; stateName shows UPDATING while Atlas applies the change. Use this to change an existing cluster. Use mongodb_cluster_create for a new one, or mongodb_flex_cluster_update for Flex clusters. Requires a project id and cluster name from mongodb_clusters_list.

Parameters

Name

Type

Required

Description

cluster_name

string

Required

Name of the cluster. Get it from mongodb_clusters_list. Example: Cluster0

group_id

string

Required

Unique 24-hexadecimal id of the Atlas project (Atlas calls it groupId). Get it from mongodb_projects_list. Example: 32b6e34b3d91647abb20e7b8

backup_enabled

boolean

Optional

Flag that indicates whether the cluster can perform backups. If set to true, the cluster can perform backups. You must set this value to true for NVMe clusters. Atlas defaults to false. Example: true

mongodb_major_version

string

Optional

MongoDB major version of the cluster. Set to the binary major version. On creation: Choose from the available versions of MongoDB, or leave unspecified for the current recommended default in the MongoDB Cloud platform. Example: 8.0

paused

boolean

Optional

Flag that indicates whether the cluster is paused. Example: false

pit_enabled

boolean

Optional

Flag that indicates whether the cluster uses continuous cloud backups. Atlas Infinite clusters always use continuous cloud backups. Atlas sets this flag to true on create and update for Atlas Infinite clusters and ignores a requested value of false. Example: false

replication_specs

array

Optional

List of settings that configure your cluster regions. This array has one object per shard representing node configurations in each shard. For replica sets there is only one object representing node configurations. Example: [{"regionConfigs": [{"providerName": "AWS", "regionName": "US_EAST_1", "priority": 7, "electableSpecs": {"instanceSize": "M10", "nodeCount": 3}}]}]

tags

array

Optional

List that contains key-value pairs between 1 to 255 characters in length for tagging and categorizing the cluster. Example: [{"key": "environment", "value": "production"}]

termination_protection_enabled

boolean

Optional

Flag that indicates whether termination protection is enabled on the cluster. If set to true, MongoDB Cloud won't delete the cluster. If set to false, MongoDB Cloud will delete the cluster. Atlas defaults to false. Example: false

version_release_system

string

Optional

Method by which the cluster maintains the MongoDB versions. If value is CONTINUOUS, you must not specify mongoDBMajorVersion. Atlas defaults to LTS. Allowed values: LTS, CONTINUOUS. Example: LTS. One of: `LTS`, `CONTINUOUS`.

mongodb_clusters_list

List clusters

mongodb_custom_role_create

Create custom role

mongodb_custom_role_get

Get custom role

mongodb_custom_roles_list

List custom roles

mongodb_database_user_delete

Delete database user

mongodb_flex_cluster_create

Create flex cluster

mongodb_flex_cluster_get

Get flex cluster

mongodb_flex_cluster_update

Update flex cluster

mongodb_flex_clusters_list

List flex clusters

mongodb_online_archive_get

Get online archive

mongodb_organization_events_list

List organization events

mongodb_organizations_list

List organizations

mongodb_process_measurements_get

Get process measurements

mongodb_processes_list

List processes

mongodb_project_get

Get project

mongodb_projects_list

List projects

mongodb_schema_advice_get

Get schema advice

mongodb_search_index_create

Create search index

mongodb_slow_queries_list

List slow queries

List the slow-query log lines Atlas recorded for one MongoDB process. Returns slowQueries entries, each with the namespace and the raw log line; the include flags add parsed metrics, opType, and replicaState. Use this to inspect individual slow operations. Use mongodb_suggested_indexes_list for index recommendations. Requires a project id and process id from mongodb_processes_list.

Parameters

Name

Type

Required

Description

group_id

string

Required

Unique 24-hexadecimal id of the Atlas project (Atlas calls it groupId). Get it from mongodb_projects_list. Example: 32b6e34b3d91647abb20e7b8

process_id

string

Required

MongoDB process id in hostname:port form, as returned in the id field of mongodb_processes_list. Example: atlas-abc123-shard-00-00.x1y2z.mongodb.net:27017

duration

integer

Optional

Length of time expressed during which the query finds slow queries among the managed namespaces in the cluster. This parameter expresses its value in milliseconds. - If you don't specify the since parameter, the endpoint returns data covering the duration before the current time. - If you specify neither the duration nor since parameters, the endpoint returns data from the previous 24 hours. Example: 3600000

include_metrics

boolean

Optional

Whether or not to include metrics extracted from the slow query log as separate fields. Atlas defaults to false. Example: false

include_op_type

boolean

Optional

Whether or not to include the operation type (read/write/command) extracted from the slow query log as a separate field. Atlas defaults to false. Example: false

include_replica_state

boolean

Optional

Whether or not to include the replica state of the host when the slow query log was generated as a separate field. Atlas defaults to false. Example: false

max_logs

integer

Optional

Maximum number of lines from the log to return. Atlas defaults to 20000. Example: 100

namespaces

array

Optional

Namespaces from which to retrieve slow queries. A namespace consists of one database and one collection resource written as .: <database>.<collection>. To include multiple namespaces, pass the parameter multiple times delimited with an ampersand (&) between each namespace. Example: ["sample_mflix.movies"]

since

integer

Optional

Date and time from which the query retrieves the slow queries. This parameter expresses its value in the number of milliseconds that have elapsed since the UNIX epoch. - If you don't specify the duration parameter, the endpoint returns data covering from the since value and the current time. - If you specify neither the duration nor the since parameters, the endpoint returns data from the previous 24 hours. Example: 1760400000000

For more tools, view docs.

Build your Agent

Same auth pattern across LangChain, OpenAI, Anthropic, and Google ADK.

import { ScalekitClient } from "@scalekit-sdk/node";
import { createReactAgent } from "@langchain/langgraph/prebuilt";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

// MongoDB Atlas tools available to this agent
const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["mongodb"], toolNames: [
    "mongodb_clusters_list",
    "mongodb_alerts_list",
    "mongodb_suggested_indexes_list"] },
  pageSize: 100,
});

const agent = createReactAgent({ llm, tools });
await agent.invoke({ messages: [{ role: "user", content: "Check open alerts on the Production project and suggest index fixes" }] });

import OpenAI from "openai";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const openai = new OpenAI();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["mongodb"] }, pageSize: 100,
});

const res = await openai.chat.completions.create({
  model: "gpt-5",
  messages: [{ role: "user", content: "Check open alerts on the Production project and suggest index fixes" }],
  tools,
});

// Execute the tool call with the vaulted MongoDB Atlas service account credentials
await sk.tools.executeTool(res.choices[0].message.tool_calls[0], "user_123");

import Anthropic from "@anthropic-ai/sdk";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const anthropic = new Anthropic();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["mongodb"] }, pageSize: 100,
});

const msg = await anthropic.messages.create({
  model: "claude-sonnet-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Check open alerts on the Production project and suggest index fixes" }],
  tools,
});

// Tool call runs with the vaulted MongoDB Atlas service account credentials
await sk.tools.executeTool(msg.content, "user_123");

import { Agent } from "@google/adk/agents";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["mongodb"] }, pageSize: 100,
});

const agent = new Agent({
  name: "database_agent",
  model: "gemini-2.5-pro",
  instruction: "MongoDB Atlas tools available to this agent",
  tools,
});

await agent.run("Check open alerts on the Production project and suggest index fixes");

Try these prompts

Copy any prompt into your agent. Each maps directly to a MongoDB Atlas tool. Click to copy, paste into your agent, done.

Clusters and alerts

Copy the prompt

Copied

List the clusters in every Atlas project we can access.

Copy the prompt

Copied

Show open alerts on the Production project.

Copy the prompt

Copied

Acknowledge the disk usage alert until tomorrow at 9am.

Performance

Copy the prompt

Copied

Which indexes does Performance Advisor suggest for prod-cluster-0?

Copy the prompt

Copied

Show slow queries on the primary from the last hour.

Copy the prompt

Copied

Get CPU and connection metrics for the primary over the last 24 hours.

Backups and access

Copy the prompt

Copied

Take an on-demand backup snapshot of prod-cluster-0.

Copy the prompt

Copied

Add 203.0.113.10 to the Production project's IP access list.

Copy the prompt

Copied

Create a read-only database user for the analytics service.

SEE HOW AUTH WORKS

You connect MongoDB Atlas once with a service account, and Scalekit authenticates every call. Service account credentials stay vaulted, every call is scope checked, and every action is logged.

Authorize

Your user connects

MongoDB Atlas

once. We tie it to their identity and the meetings they approved — no shared bot account, no org-wide access

Who:

user ‘A’

when:

Once per user

access:

Limited to user

Store

Their

MongoDB Atlas

token lives in a vault scoped to them. User A's meetings are never reachable by an agent acting for user B, even on the same connection

vault:

encrypted

scope:

per-user

tokens:

auto-refreshed

Resolve

When your agent calls a

MongoDB Atlas

tool, we fetch the right token server-side. It never touches your agent, never appears in the LLM context, never shows up in your logs

speed:

~40ms

check:

before every call

seen by:

nobody

Audit

Every

MongoDB Atlas

tool call is logged — who triggered it, which meeting was fetched, what came back. 90 days of history, tied to the user who authorized it

history:

90 days

export:

SIEM-ready

logged:

every call

Test other agents

See the same auth pattern across other database and developer connectors.

Engineering Teams

DevOps assistant agent

Polls GitHub for failing checks and stale PRs, opens Linear issues for the ones that need work, and posts a daily digest to Slack. It acts as the engineer, not a shared service account.

Engineering Teams

Engineering standup agent

Pulls commits from GitHub and GitLab, tracks issue movement in Jira, and posts a per-engineer standup brief to Slack. Each engineer's activity is read on their own delegated OAuth.

Why Scalekit

Secure your agent's access. Connectors ship in minutes

01.

Service account secrets stay out of agent code

Pasting a MongoDB Atlas service account secret into agent code puts it one log line away from a leak. Scalekit keeps the service account credentials in its vault, authenticates each call at request time, and records which agent run made it.

// secret in code
env → prompt, logs

// scalekit
vault → call ✓

02.

Authentication is not authorization

03.

Multi-tenancy is architectural

04.

MongoDB Atlas today. Ten connectors tomorrow.

“Our agents act across Salesforce, Gong, Google Drive, and more, on behalf of every customer. Scalekit behind the scenes meant we can keep adding tools without ever rebuilding how credentials or tool calling work.”

Venu Madhav Kattagoni

Head of Engineering / Von

FAQs

Frequently Asked Questions

Does the agent use a shared service account for MongoDB Atlas?

Yes. You connect MongoDB Atlas once with a service account, and Scalekit authenticates every call. The agent reaches the Atlas organizations and projects that service account can access, and audit logs record every action the agent takes.

Where are the MongoDB Atlas service account credentials stored?

In Scalekit's managed AES-256 token vault, namespaced per tenant. Revocation is a single dashboard action. Credentials never appear in prompts, logs, or LLM context.

Can I limit what the agent does in MongoDB Atlas?

Yes. Filter by tool name in listScopedTools to expose only what you want, for example the cluster, alert, and Performance Advisor read tools without mongodb_cluster_delete or mongodb_project_delete, which permanently remove clusters and projects. Scalekit also enforces scope checks before every API call.

What happens when MongoDB Atlas access is turned off?

Disable the connection in the Scalekit dashboard and the next tool call fails closed with a clear error. Other tenants are unaffected, and the change is logged.

Does the agent query the documents in my collections?

No. The tools manage Atlas itself: clusters, backups, alerts, metrics, database users, IP access lists, and search and regular indexes. mongodb_process_databases_list and mongodb_process_database_measurements_get return database names and storage metrics, not documents.