Provider REST API

CRM

Connect Claude to Salesforce

Query records with SOQL, read and update any object through the REST API. Toolspoke puts 9 of its actions behind one MCP endpoint that Claude, Cursor, and Codex all speak.

Connection
Provider REST API
Authentication
Sign in with Salesforce
Actions exposed
9
Cost per call (typical)
1 credit
Adapter
Maintained by Toolspoke

Connected in three steps

  1. 1

    Install Salesforce

    Open the marketplace in your workspace, add Salesforce to the project your agents work in, and it appears on the gateway immediately.

  2. 2

    Connect the credential

    Authenticate with sign in with salesforce. Where to get one, and what it has to be able to reach, is the next section.

  3. 3

    Point your agent at the gateway

    Give your client one address, https://toolspoke.com/mcp. Claude Code takes it as a command, Claude and Claude Desktop add it as a custom connector, and Cursor, Codex and VS Code each read it from a config file of their own.

.mcp.json
{
  "mcpServers": {
    "toolspoke": {
      "type": "http",
      "url": "https://toolspoke.com/mcp"
    }
  }
}

One block covers every tool you have installed. Salesforce shows up in the client as soon as your policy allows it, and so does everything else you install later.

Where the address goes, per client

Claude Code

Run it in your project, then /mcp to sign in

claude mcp add --transport http toolspoke https://toolspoke.com/mcp
Claude and Claude Desktop

Settings, then Connectors, then Add custom connector

https://toolspoke.com/mcp
Cursor

~/.cursor/mcp.json, or .cursor/mcp.json for one project

{ "mcpServers": { "toolspoke": { "url": "https://toolspoke.com/mcp" } } }
Codex

~/.codex/config.toml

[mcp_servers.toolspoke]
url = "https://toolspoke.com/mcp"
VS Code

.vscode/mcp.json, or the MCP: Add Server command

{ "servers": { "toolspoke": { "type": "http", "url": "https://toolspoke.com/mcp" } } }

What Salesforce asks for

Sign in with Salesforce. You provide it once, when you install the connector. Toolspoke encrypts it at rest and decrypts it only for the length of a single call, and the gateway attaches it to the outbound request itself, so it is never part of the arguments an agent sends.

Instance URLRequired
Your org's own API host - the instance_url Salesforce shows in Setup → Company Information, or the host in the address bar once you are logged in. It differs per org, so there is no default that would work.
https://acme.my.salesforce.com
API versionOptional
The REST API version to call, with its leading v. Defaults to v62.0, which every currently supported org accepts. Raise it (v67.0 is Summer '26) only if you need something newer.
v62.0

What Claude can do in Salesforce

9 actions, each one declared and named by the connector rather than discovered at runtime. A workspace policy grants a person all of them, a hand-picked selection, everything on the read side, everything on the write side, or none.

Reads
7Reads
Writes
2Writes
Destructive
0Destructive

Reads

7

Fetches data and changes nothing.

  • soql_query

    Run a SOQL query and return the matching records. This is the main way to read anything out of Salesforce: it reaches every standard and custom object, follows relationships (Account.Name from a Contact), filters, sorts and aggregates. Name the fields you want rather than guessing - SOQL has no SELECT * - and call describe_object first when you are not sure a field exists, because a bad field name comes back as an error rather than as an empty result. Salesforce returns the first 2,000 records and sets done to false when there are more; this connector has no follow-the-cursor operation, so page with LIMIT and OFFSET in the statement itself, or narrow the WHERE clause.

  • sosl_search

    Run a SOSL search - full-text matching across several objects at once, which SOQL cannot do. Use it when you have a name, a phone number or a fragment of an email address and do not yet know which object it lives on; use soql_query when you know the object and the field. The statement carries its own structure, e.g. FIND {Acme} IN ALL FIELDS RETURNING Account(Id, Name), Contact(Id, Name, Email), Lead(Id, Name).

  • get_record

    Fetch one record by its 15 or 18 character Salesforce id, on any object. Ask for the fields you need - without `fields` Salesforce returns every field on the record, which on a customised Account or Case is thousands of tokens of mostly empty columns. Ids come from soql_query, sosl_search or list_recent_items; describe_object lists the field names this object actually has.

  • describe_object

    Return one object's full metadata: every field with its API name, label, data type, whether it is required, and the picklist values it accepts. This is the operation to call before writing anything, because a Salesforce org's schema is whatever its admins built - field names, required fields and picklist values are all local. The response is large; ask for one object at a time.

  • list_objects

    List every object in the org - standard and custom - with its API name, label and what can be done to it. Call this first when you do not know what the org calls something: a company's Accounts may be Accounts, or they may have a custom object beside them. Then call describe_object for the one you want. This is also the connection check, because it is a metadata read that touches no records.

  • list_recent_items

    Return the records the signed-in user viewed most recently, across objects, newest first. Useful as an opening move when somebody says "the opportunity I was just looking at" and gives you nothing else to go on. Returns ids and names only; pass an id to get_record for the rest.

  • get_current_user

    Return the identity this connection is signed in as: the Salesforce user id, username, display name, email address and organization id. Call it to confirm which org and which user a write would be attributed to before making one. It needs the `id` scope, which this connector's sign-in requests; list_objects is the check used at install because it needs only `api`.

Writes

2

Creates or updates something on the other side.

  • create_record

    Create one record on any object. `body` is a flat map of field API names to values, exactly as Salesforce takes it - {"Name": "Acme Corp", "Industry": "Technology"} for an Account, {"LastName": "Doe", "Email": "[email protected]", "AccountId": "001…"} for a Contact. Call describe_object first for the field names and for which of them are required; Salesforce rejects the whole create when one required field is missing and names it in the error. Returns the new record's id.

  • update_record

    Change fields on an existing record. Only the fields in `body` are written; everything else is left alone, and null clears a field. Salesforce answers a successful update with an empty 204, so read the record back with get_record if you need to see the result. This cannot delete a record - no delete operation is offered by this connector at all.

What it will not do

Enforced by the gateway rather than left to convention, which is why each of these can be stated flatly.

It cannot call anything else
The 9 actions above are the whole of it. A call to any other name is refused before it reaches Salesforce rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time.
Nothing here deletes
This connector writes to Salesforce, but nothing in it deletes or permanently alters anything.
It reaches no further than your credential
Toolspoke holds no access to Salesforce of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either.
It never hears from Salesforce
Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Salesforce. An agent has to ask.
It does not smooth over provider limits
Toolspoke does not retry, queue or back off around Salesforce's own rate limits. A call that Salesforce refuses comes back to the agent as a failed call.

Before you connect it

What can Claude do in Salesforce?

9 named actions: 7 that only read and 2 that write. They include soql_query, sosl_search and get_record. Nothing outside that list is reachable: the connector declares each operation by name rather than proxying whatever an agent asks for.

What credentials does the Salesforce connector need?

Sign in with Salesforce. The connector asks for instance url, and optionally api version. Values are encrypted at rest and attached to the outbound request by the gateway, so they are never part of the arguments an agent sends and never reach the audit log.

Does the Salesforce connector work with Cursor and Codex, or only Claude?

Any client that speaks MCP, and every one of them gets the same 9 actions. There is a single address, https://toolspoke.com/mcp. Claude Code adds it with claude mcp add --transport http, Claude and Claude Desktop take it as a custom connector in settings, Cursor reads it from .cursor/mcp.json, Codex from ~/.codex/config.toml, and VS Code from .vscode/mcp.json. Each of them signs in to the gateway itself, so there is no key to paste.

What does the Salesforce connector not do?

The 9 actions above are the whole of it. A call to any other name is refused before it reaches Salesforce rather than forwarded on, and connecting your account does not add to the list: it is fixed by the connector, not discovered at run time. This connector writes to Salesforce, but nothing in it deletes or permanently alters anything. Toolspoke holds no access to Salesforce of its own. Every call carries the credential you stored and nothing besides, so whatever that credential cannot reach, this connector cannot reach either. Nothing is pushed to it. There is no webhook, no subscription and no polling, so this connector cannot notice by itself that something changed in Salesforce. An agent has to ask. Toolspoke does not retry, queue or back off around Salesforce's own rate limits. A call that Salesforce refuses comes back to the agent as a failed call.

Can I limit which actions an agent can call?

Yes, in two places. The project switches Salesforce's actions on and off one at a time, for everyone in the project at once, and the screen groups them by read, write and destructive so turning off everything that deletes is one click. An individual agent key can then be narrowed further, to particular toolkits in a project and to particular actions in a toolkit. Whatever it was granted, a key never reaches a project its owner cannot.

What gets recorded when an agent calls Salesforce?

Every attempt, with the agent that made it and the person that agent belongs to, the full request payload, the response payload, the status, the duration, and the credits spent. Values whose key names a secret are masked out before the record is shown to anyone. An operation the connector marks as not retained never has its response body written at all, so the gateway keeps no second copy of what was read.