Skip to main content

Scispot MCP Server (Beta)

Control Your Lab with AI — From Any LLM

S
Written by Satya Singh
Updated today

Beta Access Required · The Scispot MCP Server is currently in private beta. To get access, contact our Customer Success team at [email protected]

What Is This?

The Scispot MCP (Model Context Protocol) Server lets you use any AI assistant — Claude, ChatGPT, Gemini, etc — to control your Scispot lab workspace conversationally.

Ask questions, update data, create experiments, and analyze results using plain English. No coding required.

Key Features:

✅ Works with Claude Desktop, ChatGPT, and Gemini

✅ Secure OAuth authentication

✅ 17 powerful tools for labsheets, experiments, samples, and protocols

✅ Read and write access (with your existing permissions)

Quick Start

Prerequisites

  • An active Scispot account

  • Claude Desktop, ChatGPT Plus, Gemini Advanced, or Cursor installed

  • Beta access approval from [email protected]

  • API Key access, instructions for generating API key can be found here

Setup Instructions

Setup Scispot MCP connector on the web (Claude.ai)

  1. Go to the settings page and click on the connectors setting or visit this link to directly access the page.
    ​

  2. Click on add custom connector

  3. Expand the advanced details and fill in the following details:

    1. Name: Scispot

    2. Remote MCP server URL: https://mcp.scispot.io/

    3. OAuth Client ID: claude-ai-client

    4. OAuth Client Secret: Paste your Scispot API Key here

  4. You should now see the Scispot connector in your list of connectors

Setup Scispot MCP connector on Claude Code (Terminal)

  1. If you have previously installed the scispot-mcp, remove it using the command:
    ​

    claude mcp remove scispot-mcp

  2. Now add the mcp connector to claude using this command, replacing YOUR_SCISPOT_API_KEY with the Scispot API key.
    ​

    claude mcp add scispot-mcp https://mcp.scispot.io/ --transport http --header "Authorization: Bearer YOUR_SCISPOT_API_KEY"

  3. Now start up claude code and use the /mcp command to access your list of mcp servers.

  4. You should see the scispot-mcp as connected, if there is any issues you can also manually reconnect to the mcp here.

Contact [email protected]

What You Can Do

Ask Questions in Plain English

  • "Show me all samples in the -80°C freezer"

  • "What experiments did Sarah run last week?"

  • "Which assays have TAT over 48 hours?"

  • "Find all samples where Status=Active AND Concentration > 10mg/mL"

  • "Look at experiment EXP-042, research related publications, and suggest protocol updates"

Automate Routine Tasks

  • Update hundreds of labsheet rows with a single prompt

  • Generate experiment documentation from voice notes

  • Cross-reference protocols, materials, and results

  • Track sample locations across your facility

  • Bulk-add QC results from instrument outputs

Get Instant Insights

  • Search across all labsheets, experiments, and protocols simultaneously

  • Filter and analyze data without writing queries

  • Link experimental data to inventory and documentation

  • Access your entire lab knowledge base conversationally

For Every Lab Role

🧪 Wet Lab Scientists

Spend less time on data entry, more time at the bench.

  • Voice-to-data: "Add 12 new mouse samples to the breeding colony tracker with these IDs..."

  • Quick lookups: "Where is the anti-FLAG antibody stored?" → Get location, lot number, and expiration instantly

  • Protocol access: "Show me the latest cell culture protocol for HEK293 cells"

  • Documentation: "Create an experiment note for today's Western blot with these observations..."

No coding required. Just ask in natural language.

💻 Computational Scientists & Bioinformaticians

Accelerate data analysis workflows.

  • Batch operations: Update metadata for thousands of sequencing runs

  • Data integration: Cross-reference NGS results with sample metadata from labsheets

  • Automated reporting: Generate sample manifests, QC reports, and data summaries

  • Schema exploration: Discover available data fields and relationships without SQL

API-like power with conversational simplicity.

📊 Lab Managers & Operations

Stay on top of inventory, compliance, and throughput.

  • Inventory monitoring: "Which reagents expire in the next 30 days?"

  • Capacity planning: "How many samples are pending QC review?"

  • Team productivity: "List all experiments created"

Real-time visibility without dashboards or reports.

🔬 Principal Investigators

Oversee research progress conversationally.

  • Project overview: "Summarize all experiments for the kinase inhibitor project"

  • Resource allocation: "Which equipment has been most utilized this quarter?"

  • Data exploration: "Find all assays with IC50 values below 10nM"

  • Collaboration: "Show me protocols shared between the Boston and Cambridge sites"

Strategic insights without digging through folders.

AI Platform Support

Compatible AI Assistants

The Scispot MCP Server works with all major AI platforms that support the Model Context Protocol:

Platform

Setup Method

Best For

Claude Desktop

OAuth (one-click)

Complex reasoning, multi-step workflows

ChatGPT Plus

Third-party integration

Broad familiarity, mobile access

Gemini Advanced

Extension marketplace

Fast responses, multimodal queries

Cursor / VS Code

MCP extension

Developers, computational biologists

You choose the AI model that fits your budget and performance needs. Switch between models seamlessly without changing your Scispot integration.

Capabilities: 17 Powerful Tools

Labsheets (7 tools)

Your structured lab data, accessible through conversation.

Tool

What It Does

Example Use Case

list_labsheets

Discover all labsheets in your workspace

"What labsheets do I have access to?"

get_labsheet_schema

View column definitions and data types

"Show me the fields in the Sample Tracking sheet"

search_labsheet_rows

Filter and query rows with AND/OR logic

"Find all samples where Status=Active AND Location=Freezer 3"

get_labsheet_row

Retrieve a single row by ID

"Get the details for sample UUID abc-123"

add_labsheet_rows

Bulk insert new rows

"Add these 100 compound screening results"

update_labsheet_rows

Modify existing data

"Update the concentration for these 10 samples"

list_labsheet_folders

Browse folder hierarchy

"Show me labsheets in the Genomics folder"

Electronic Lab Notebook (5 tools)

Manage experiments, protocols, and documentation.

Tool

What It Does

Example Use Case

list_experiments

Browse experiments by location

"Show me experiments in the Q1 2026 folder"

get_experiment

Read full experiment content

"Open the CRISPR screen from Feb 15"

create_experiment_stub

Start a new experiment entry

"Create an experiment called 'Kinase Assay Optimization'"

append_to_experiment

Add notes, observations, results

"Add these Western blot observations to experiment EXP-042"

list_protocols

Find standard operating procedures

"List all protocols in the Cell Culture folder"

Sample & Inventory Management (3 tools)

Track materials, samples, and storage.

Tool

What It Does

Example Use Case

get_manifest

View plate/box/rack contents

"Show me what's in plate PLT-001"

list_locations

Browse storage hierarchy

"List all freezers in the lab"

list_labflow_samples

Get samples in a workflow

"Show me all samples in Labflow FLOW-43"

Cross-Linking & Documentation (2 tools)

Connect data across your lab ecosystem.

Tool

What It Does

Example Use Case

get_protocol

Read protocol details

"Show me the DNA extraction protocol"

get_documentation

Access reference documentation

"Open the equipment calibration logs"

Security & Compliance

Enterprise-Grade Security

Your data security is our top priority. The Scispot MCP Server is built with multiple layers of protection:

Authentication & Authorization

  • OAuth 2.0 + PKCE: Industry-standard authentication protocol used by Google, Microsoft, and major financial institutions

  • Personal access tokens: Each user generates their own access token — no shared credentials

  • Token expiration: Automatic token lifecycle management

  • Scope-based permissions: Tools only access data you explicitly authorize

Data Protection

  • Encrypted transit: All data transmitted over HTTPS/TLS 1.3

  • No training data: Your proprietary lab data is never used to train AI models

  • AWS infrastructure: Hosted on AWS with SOC 2, HIPAA-eligible, and ISO 27001 compliance

  • Audit logging: All API calls are logged with user identity, timestamp, and action details

What the AI Can and Cannot Do

✅ The AI assistant can:

  • Read data you have permission to view

  • Create and modify labsheet rows

  • Create and append to experiments

  • Search across your accessible data

❌ The AI assistant cannot:

  • Access data outside your Scispot account

  • Delete labsheets or experiments (read/create/update only)

  • Change user permissions or system settings

  • Export data outside Scispot without your explicit command

  • Execute arbitrary code or system commands

You remain in control. You can review every action before it executes, and all changes are traceable in Scispot's audit log.

HIPAA & Regulatory Compliance

Important: Scispot infrastructure supports HIPAA workloads and FDA 21 CFR Part 11 requirements. Contact [email protected] for a Business Associate Agreement (BAA) if you handle PHI.

Customer responsibility: When using external AI models (Claude, ChatGPT, Gemini), you are responsible for ensuring compliance with your data handling policies. The MCP connection provides secure access, but data sent to external AI providers is governed by their terms of service.

For regulated environments, we recommend:

  • Using Claude for Enterprise or ChatGPT Enterprise with BAAs

  • Reviewing your organization's AI usage policies

  • Training staff on what data can be shared with AI assistants

Beta Limitations & Roadmap

Current Limitations (Beta)

  • Rate limits: 60 requests per minute per user

  • Batch operations: Limited to 1,000 rows per request

  • Image handling: Thumbnails not yet available via MCP (coming soon)

  • Advanced linking: Labsheet-to-protocol/documentation linking coming in v1.0

What's Coming Next

v1.0 (April 2026)

  • Image thumbnail support (get_image_thumbnail)

  • Cross-linking tool (link_labsheet_to_protocols)

  • Comprehensive audit trail for MCP requests

  • Improved batch performance (10K rows)

FAQ

General

Q: Do I need to know how to code?
A: No. Just ask in natural language — the AI handles everything.

Q: Which AI platform is best?
A: It depends on your needs. Claude excels at complex reasoning, ChatGPT has broad adoption, and Gemini offers competitive pricing. Many teams use multiple models.

Q: Can I use this on mobile?
A: Yes, if your AI platform has a mobile app (Claude and ChatGPT both do).

Q: Does this work offline?
A: No. The integration requires internet access to query your Scispot workspace.

Security

Q: Can I audit what the AI does?
A: Yes. All API calls are logged in Scispot with user, timestamp, and action details. Go to Settings → Audit Log to review.

Q: Is this HIPAA compliant?
A: Scispot infrastructure supports HIPAA workloads. Contact us for a Business Associate Agreement (BAA) if you handle PHI. Important: You are responsible for compliance when data is sent to external AI models (Claude, ChatGPT, etc.). Use enterprise AI plans with BAAs for regulated data.

Q: What if someone unauthorized gets my OAuth token?
A: Tokens expire automatically and are scoped to your permissions. You can revoke access anytime in Settings → Connected Apps. All actions are logged.

Technical

Q: What if the AI makes a mistake?
A: You can undo changes via Scispot's revision history. For critical operations, we recommend reviewing AI-generated changes before confirming.

Q: Can I integrate this with my own scripts?
A: Yes. The MCP server can be called from any MCP-compatible client, including custom applications. See the technical documentation for details.

Q: What's the difference between this and the Scispot API?
A: The MCP server is designed for conversational AI access. The REST API is for programmatic integrations. Both use the same underlying data and permissions.

Q: Can I use this in a regulated GxP environment?
A: Yes, with caveats. Scispot supports 21 CFR Part 11, but AI-generated content must be reviewed by qualified personnel before signing off. Contact [email protected] for validation support packages.

Get Early Access

The Scispot MCP Server is currently in private beta.

To request access:

  1. Email [email protected] with the subject line: "MCP Beta Access Request"

  2. Include:

    • Your Scispot workspace name

    • Which AI platform(s) you plan to use (Claude, ChatGPT, Gemini, Cursor)

    • Your primary use case (data entry, analysis, experiment documentation, etc.)

  3. Our team will enable MCP access for your account within 1-2 weeks depending upon your plan

  4. You'll receive setup instructions and your OAuth client credentials

Questions? Email [email protected] or book a demo at scispot.com/demo

Did this answer your question?