Salesforce MCP. Manage your entire sales pipeline through natural conversation.
Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Salesforce MCP Server lets your AI agent manage everything in Salesforce. Run complex SOQL queries, list accounts and contacts, or create new records instantly.
You can perform full CRUD operations on Accounts, Opportunities, Leads, and more by just talking to it. It connects your entire sales pipeline directly to your AI client.
What your AI agents can do
Create record
Creates a brand new Salesforce record when you provide JSON fields as a string.
Delete record
Permanently deletes a specified Salesforce record; this action cannot be undone.
Describe object
Retrieves metadata about a Salesforce object so you can understand its available fields and structure.
The global_search tool runs a search that pulls matches from Accounts, Contacts, Leads, and other relevant objects simultaneously.
Use the soql_query tool to execute any custom query against Salesforce records for granular filtering.
The create_record tool lets your agent build a brand-new record (like an Account or Contact) using specified JSON data.
You can modify any existing Salesforce record by calling update_record, specifying the object, ID, and new field values.
The get_record tool pulls all data for a single, known Salesforce record type (like an Account or Opportunity) by its unique ID.
Ask AI about this MCP
Supported MCP Clients
Waiting for input…
Salesforce: 12 Tools for CRM Data Operations
These tools let you automate reading, creating, updating, and deleting records in Salesforce using your AI agent.
019d7602create record
Creates a brand new Salesforce record when you provide JSON fields as a string.
019d7602delete record
Permanently deletes a specified Salesforce record; this action cannot be undone.
019d7602describe object
Retrieves metadata about a Salesforce object so you can understand its available fields and structure.
019d7602get record
Pulls all data for one specific record by knowing its SObject type (e.g., Account) and its unique ID.
019d7602global search
Searches across every major Salesforce object—Accounts, Contacts, Leads, etc.—to find matches simultaneously.
019d7602list accounts
Retrieves a list of all existing accounts within your connected Salesforce instance.
019d7602list contacts
Fetches a list of contacts from the specified Salesforce dataset.
019d7602list opportunities
Lists all active sales opportunities, showing their current stage and value.
019d7602list reports
Gets a list of report types available in your Salesforce environment so you know what reports exist.
019d7602run report
Executes a specific, named Salesforce report and returns the data results directly.
019d7602soql query
Allows you to execute any custom SOQL query, giving you precise control over which fields and records are returned.
019d7602update record
Modifies an existing Salesforce record by providing the object type, ID, and a set of new JSON field values.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Salesforce, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Look, you've got a CRM full of data—accounts, leads, opportunities—and talking to it is usually a headache. This MCP Server changes that. It lets your AI client talk directly to Salesforce, giving it the ability to manage everything in your sales pipeline using plain conversation. You get full control over reading and writing every single piece of data.
When you need to know what's going on across the board, you don't want to run ten different reports. That's why we got global_search; it hits Accounts, Contacts, Leads, and all those other major objects at once, pulling back matches from every place simultaneously. If that ain't enough searching power, you can use soql_query.
This tool lets you write any custom SOQL query you want, giving you surgical precision over which fields and exactly which records come back. You dictate the filter, period.
Need to build out a new client record from scratch? You gotta use create_record. Just feed it the JSON structure for what you need—say, an Account or a Contact—and your agent builds that brand-new Salesforce entry instantly. But if you just got bad data and realize you don't need it anymore, delete_record handles the cleanup.
Be careful with this one; it permanently deletes stuff, so make sure what you tell it to delete is actually garbage.
When you only care about one specific record—like that big opportunity for Acme Corp—you use get_record. You know the SObject type and you have the unique ID? This tool pulls all the data associated with that single record. Wanna see what kinds of records exist in your environment? Run describe_object; it spits out the metadata, showing you all the available fields and the object's structure so you know exactly what to expect.
If you just want a quick overview instead of digging into IDs, there are list tools. You can use list_accounts to pull a complete roster of every account in your system, or run list_contacts to grab all the contacts from that dataset. For sales tracking, list_opportunities shows you a rundown of all active opportunities, letting you see their current stage and dollar value right off the bat.
And if you wanna know what kind of reports are even possible in your system, run list_reports to get that master list of available report types.
Now for modifying data—this is where it gets powerful. If a record exists but some fields are wrong or outdated, don't sweat it. You use update_record. You specify the object type, give it the ID, and then you feed in the new JSON values you want to stick. It modifies that existing data point for you.
For deep dives, if a specific report is needed, running run_report executes that named report, and your agent gets those results right back. Remember, when all this feels too complex for direct queries, you always have the soql_query fallback to keep things granular.
You've got everything: from getting a list of available reports with list_reports, to executing that report data using run_report. You can manage full CRUD operations—Create, Read, Update, Delete—on every SObject type. Whether you need to check the structure first with describe_object or pull specific details via get_record, your AI client handles it all natively within the Salesforce environment.
How Salesforce MCP Works
- 1 You tell your AI agent what you need—for example, 'Show me high-value opportunities in the Technology sector.'
- 2 The agent recognizes this as a data request and selects the appropriate tool, like
soql_query, passing it the required parameters. - 3 The MCP Server executes the query against Salesforce and sends the structured result (e.g., a list of records) back to your AI client for you to review.
The bottom line is that your agent handles the API calls, so you just ask it what data you need in plain English.
Who Is Salesforce MCP For?
This is for anyone who spends time in Salesforce but hates clicking through multiple tabs. If you're an Account Manager tired of copying and pasting pipeline status updates, or a Business Analyst who needs to cross-reference report data with raw contact records, this saves you hours.
Uses list_reports and run_report to audit sales metrics across different regions without manual CSV exports.
Calls create_record or update_record immediately after a client call to ensure all new contacts and opportunity stages are logged instantly.
Runs complex queries using soql_query to pull datasets that need manual joining outside of Salesforce, making the data available for deeper analysis.
What Changes When You Connect
- Stop manually checking dashboards. Use
list_accountsorlist_opportunitiesto pull lists of records instantly, allowing your agent to process hundreds of records in seconds. - No more guesswork on fields. Running
describe_objectfirst lets you confirm the exact data structure before attempting any write operation withcreate_recordorupdate_record. - Need a specific cross-section? Instead of building complex report filters, just use
soql_query. This tool gives you direct SQL control over your Salesforce data. - It handles all the tedious linking. The
global_searchtool doesn't force you to pick one object; it searches Accounts, Contacts, and Leads simultaneously for a single search term. - Data integrity is key. Use dedicated listing tools (
list_contacts,list_accounts) first, then useget_recordfor specific deep dives on crucial records.
Real-World Use Cases
Quickly auditing a target account's history
A manager needs to know everything about Acme Corp. Instead of clicking into the Account record, they ask their agent. The agent runs get_record on the Account ID and then uses soql_query to pull all related Contacts and Opportunities simultaneously, giving a complete snapshot.
Batch updating deal stages after a meeting
The sales team just finished 10 meetings. They tell their agent: 'Update these ten opportunity IDs.' The agent runs update_record for each ID, changing the stage and adding notes in one go, saving hours of manual CRM entry.
Identifying missing leads from a campaign
The marketing team ran a campaign and needs to find all new contacts that haven't been assigned an account. They ask the agent to run list_contacts and filter for unassigned records, which immediately surfaces the list for follow-up.
Generating a full report on Q3 performance
Instead of navigating the Reports tab, a user asks their agent: 'Run the Quarterly Performance Report.' The agent selects and executes run_report using the correct ID found via list_reports, delivering the data table immediately.
The Tradeoffs
Trying to list everything at once
Asking the AI agent, 'Just tell me all the accounts and contacts.' This vague prompt forces the agent to guess which relationship data you actually need.
→
Be specific. Start by calling list_accounts if you only care about companies, or use global_search when you are looking for a person's name that might be listed as a contact or a lead.
Over-relying on vague search terms
Searching 'high revenue deals in Europe' using only the general chat prompt. This often fails because it doesn't know which object to filter against.
→
Use soql_query for complex filtering. Write a query like: SELECT Id, Name FROM Opportunity WHERE CloseDate = LAST_N_DAYS:90 AND Amount > 50000.
Ignoring the update process
Telling the agent to 'change this record' without specifying which fields or what their new values are. The system can’t guess your intent.
→
Always provide parameters for update_record. You must specify: 1) The SObject Type, 2) The Record ID, and 3) The exact JSON keys/values you want to change.
When It Fits, When It Doesn't
Use this MCP Server if your job requires reading or writing structured data in Salesforce. This is a full-stack CRM data tool.
Prioritize: Use list_* tools (e.g., list_accounts) when you just need a general overview of all records for an object type, as these are optimized calls. Always use soql_query when the list view isn't enough and you need to filter by multiple criteria or join different objects.
Don't Use This If: Your goal is simple data visualization (use a dedicated BI tool) or if your process involves non-structured documents (you need a document management system). You don't use this if the only thing you do in Salesforce is click buttons that run workflows; these tools require direct data manipulation.
Bottom line: If you can describe it as 'reading a filtered list of objects' or 'updating specific fields on an existing record,' then this server handles it.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Salesforce. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 12 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Pulling reports and checking records shouldn't feel like navigating a labyrinth.
Today, if you need to check the status of all opportunities over $50k in Europe, you open the Opportunity tab. You filter by amount. Then you run a report. If you want to narrow that down further—say, only those with 'Proposal' stage—you have to add another filter. This process takes five clicks and three seconds of mental fatigue.
With this MCP Server, you simply ask your agent: 'Show me all open opportunities worth over $50,000 in Europe.' The agent runs the required `soql_query` or `run_report`, pulling the final, filtered data table immediately. You get the answer without touching a single dashboard filter.
Using Salesforce MCP Server: Control your entire sales record lifecycle.
Manual tasks that vanish include: navigating between Accounts and Contacts to manually link records; running reports, exporting them to CSV, then importing the data back into a tracking sheet. Every step is friction.
Now, you can tell your agent, 'Create a new contact for this lead ID and assign it to Account X.' The agent runs `create_record` and automatically links everything correctly. You just state the goal; the server handles the complexity.
Common Questions About Salesforce MCP
How do I query custom fields using the soql_query tool? +
You include the field API name in your SELECT statement, just like any other field. For instance: SELECT Id, Name, Custom_Field__c FROM Account WHERE.... Always check object metadata first using describe_object.
Can I list all accounts and then update them with one command? +
Not in a single call. You must use list_accounts to get the IDs, and then pass those IDs individually or in batches to the update_record tool for modification.
What is the difference between global_search and soql_query? +
global_search finds a match across many different objects (Accounts, Contacts, Leads). soql_query lets you target specific object types and build complex filters using SOQL syntax.
Do I need to know the exact ID to use get_record? +
Yes. The get_record tool requires two pieces of information: the SObject Type (e.g., 'Account') and the unique record ID, which you must supply.
How do I use the `describe_object` tool to figure out available fields? +
It gives you the full metadata schema for an object. This is useful when you need to know what fields and data types exist on a Salesforce object before running queries or updates.
What format do I need to use when calling the `create_record` tool? +
You must provide JSON fields as a string. The payload needs to match the object's expected data types and include any mandatory fields required for the record to save.
Does the `delete_record` tool offer any confirmation steps? +
No, this action is irreversible. You must be certain of the exact SObject type and record ID before executing this tool, as Salesforce deletions are final.
Do I have to use `list_reports` before running a report with `run_report`? +
Yes, you do. You first run list_reports to get the specific Report ID required by the run_report function. This ensures you are pulling data from the correct source.
Can I run custom SOQL queries? +
Yes! The soql_query tool accepts any valid SOQL string. Example: SELECT Id, Name, Amount FROM Opportunity WHERE StageName = 'Closed Won' LIMIT 20.
Can I create records in any object? +
Yes! Use create_record with the SObject type (Account, Contact, Lead, etc.) and a JSON string of fields.
How do I search across all Salesforce data? +
Use global_search with your search text. It uses SOSL to find matches across Accounts, Contacts, Leads, and all other objects simultaneously.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
NVIDIA API Catalog
Cloud Engine proxy running native foundational completions natively utilizing active Nemotron and Llama3 architectures.
Microsoft Ads
Connect Microsoft Ads to any AI agent via MCP.
HubSpot CRM (Full)
Manage contacts, companies, deals, tickets, notes, owners, and pipelines — full access to your HubSpot CRM through natural conversation.
You might also like
Ahrefs
Professional SEO intelligence — audit backlinks, keywords, and domain health via AI.
Hygraph (Headless CMS)
Manage headless content via Hygraph — execute GraphQL queries, introspect schemas, and manage documents.
Churnkey
Manage customer retention and cancellation flows via Churnkey — track sessions, analyze exit surveys, and automate recovery contacts directly from any AI agent.