Skip to main content

Spryng Analysis MCP

Connect your preferred assistant to the Spryng studies you can access. This package provides a local Model Context Protocol (MCP) relay that forwards read-only requests to your organization's hosted Spryng Analysis service.

You need Python 3.10 or later, a desktop MCP client that supports local stdio servers, and a Spryng organization with Analysis MCP enabled. Installing this package does not enable the service or change your study permissions.

1. Get your connection details

In Spryng, open your study's Analysis → Connect your assistant. You can also open Organization settings → Analysis MCP at /organization/<organization-slug>/settings/analysis-mcp-tokens on your Spryng app host.

Choose your assistant, create a personal token, and copy the Analysis MCP URL shown there. The token begins with smcpa_ and is shown once. Keep it private. If setup is unavailable, ask your organization administrator to check access and enablement.

2. Install the relay

Choose one installation method. These shell examples are for macOS and Linux.

pipx

pipx install 'spryng-analysis-mcp==0.2.0'
spryng-analysis-mcp --help
command -v spryng-analysis-mcp

uv

uv tool install 'spryng-analysis-mcp==0.2.0'
spryng-analysis-mcp --help
command -v spryng-analysis-mcp

If the installed command is not found, follow your installer's PATH instructions and reopen your terminal.

Python virtual environment

python3 -m venv "$HOME/.local/spryng-analysis-mcp"
"$HOME/.local/spryng-analysis-mcp/bin/python" -m pip install 'spryng-analysis-mcp==0.2.0'
"$HOME/.local/spryng-analysis-mcp/bin/spryng-analysis-mcp" --help
printf '%s\n' "$HOME/.local/spryng-analysis-mcp/bin/spryng-analysis-mcp"

Copy the full executable path printed by your chosen method for the next step. On Windows, use the installed executable's full Windows path in your client's configuration; the shell commands above use Unix paths.

3. Add Spryng to your assistant

Open your assistant's MCP settings and add a local stdio server. For clients that accept an mcpServers JSON configuration:

{
  "mcpServers": {
    "spryng-analysis": {
      "command": "/absolute/path/to/spryng-analysis-mcp",
      "args": [
        "--url",
        "https://app.spryng.io/api/organizations/YOUR_ORGANIZATION/mcp/analysis/rpc/"
      ],
      "env": {
        "SPRYNG_ANALYSIS_MCP_TOKEN": "YOUR_PERSONAL_ANALYSIS_TOKEN"
      }
    }
  }
}

Replace command with the full executable path from installation, the URL with the exact URL from Spryng, and the token placeholder with your personal token. Do not use ~ or shell variables in command: desktop clients may not expand them. In Windows JSON, escape each path backslash as \\.

Use your client's secret storage if available. Otherwise, protect the local configuration file containing your token and never share it. Restart or reload the MCP connection after saving. Configuration formats vary by client.

4. Try a question

Ask your assistant:

List the Spryng studies I can access. For one study, describe its questions and the statistics available for each.

Then:

Show a supported statistic for one question, explain any missing-answer or privacy limits, and identify the study and question behind it.

Successful tool discovery and an answer grounded in returned study data confirm the assistant connection. Spryng's Check Spryng readiness checks server readiness; it does not prove that an external assistant is connected.

For the virtual-environment installation above, you can also check server connectivity from a terminal:

export SPRYNG_ANALYSIS_MCP_URL='YOUR_ANALYSIS_MCP_URL'
export SPRYNG_ANALYSIS_MCP_TOKEN='YOUR_PERSONAL_ANALYSIS_TOKEN'
"$HOME/.local/spryng-analysis-mcp/bin/python" -m spryng_analysis_mcp.connection_check

Exit code 0 means MCP initialization and tool discovery succeeded. It does not test a model's reasoning. With pipx or uv, verify discovery in your assistant; the package is isolated from your system Python.

Where data goes

Component What it does
Hosted Spryng Stores study data and enforces authorization, redaction, and privacy limits. Requests still reach Spryng even when your model runs locally.
This relay Runs on your computer and forwards calls to the configured Spryng Analysis URL. It does not include or start an AI model.
Local model Requires a separately configured model runtime and a client capable of calling MCP tools. Model compatibility and answer quality depend on that setup.
Cloud assistant May send prompts and returned study data to its model provider. Check your organization's policy and that provider's settings before connecting.
Institution or regional setup Requires an administrator-approved model endpoint and client configuration. Installing this relay does not provision a regional service, certify residency, or isolate a model's network traffic.

The relay does not choose a model or fall back to a cloud model. It sends the token only to the configured, validated Analysis URL, rejects HTTP redirects, ignores environment proxy settings, and does not persist the token itself. Your assistant's configuration, conversation history, and model runtime have their own storage and network behavior.

Available tools and statistics come from your Spryng server. Do not treat a missing answer as a refusal or average a three-part answer as if it were a slider. Review source evidence and the usage rules returned with statistics before relying on an assistant's interpretation.

Revoke access or troubleshoot

In Organization settings → Analysis MCP → Manage tokens, revoke a token to deny subsequent requests using it. Also remove it from your assistant's configuration.

Symptom What to check
Executable not found Use the full installed executable path in command, rather than relying on the desktop app's PATH.
Missing SPRYNG_ANALYSIS_MCP_TOKEN Put the token in the MCP client's server environment or secret settings. A terminal export alone may not reach a desktop app.
401 or 403 Check for a revoked token, the wrong organization URL, or missing access. Create a replacement token only if needed.
Endpoint unavailable Check that Analysis MCP is enabled for your organization and that you copied its current URL.
Wrong endpoint or redirect Use the exact HTTPS Analysis URL ending in /mcp/analysis/rpc/; board MCP and story-collection URLs are different services.
Tools appear but answers are unsupported Confirm that your client and model support tool calls. Ask for source evidence and compare the answer with the returned data.

Contact support@spryng.io with the package version, client name, and error message. Remove tokens and participant data before sharing diagnostics. Learn more at Spryng.

Offline read-only file mode (v0.2.0)

When Spryng is unreachable, you can analyze an authorized portable study export (--local-study-bundle study.spryng-study.json) or an Analysis workspace CSV export plus a versioned JSON sidecar you maintain (--local-file-csv with --local-file-schema). The three modes, including --url, are mutually exclusive. File mode is read-only, uses no token, environment credential or URL, and never falls back to the network.

File mode is a preview for macOS and Linux. It requires a POSIX safe file open (O_NOFOLLOW with non-blocking regular-file checks). Native Windows is not supported for file mode: it fails closed with a load error and has no fallback reader. Hosted --url mode on Windows is unchanged.

The privacy floor is a safeguard of this connector, not a claim that the file owner cannot edit the export. A bundle's snapshot_id is an integrity digest, not proof of origin.

See the guide file docs/local-file-mode.md in the package source and the synthetic example under docs/examples/.

License

This connector is licensed under the Apache License, Version 2.0. Copyright 2026 ScrumDo LLC. The full LICENSE and attribution NOTICE are included in the distribution.

The license does not grant trademark rights, except for the reasonable and customary use needed to describe the origin of the work and reproduce the NOTICE, as provided in Section 6. This license covers the connector; it does not grant access to hosted Spryng services.

Metadata

Release files for spryng-analysis-mcp 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for spryng-analysis-mcp 0.2.0
File Size Uploaded Provenance Actions
spryng_analysis_mcp-0.2.0.tar.gz 80.5 kB Without provenance Details for spryng_analysis_mcp-0.2.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for spryng-analysis-mcp 0.2.0
File Interpreter ABI Platform Actions
spryng_analysis_mcp-0.2.0-py3-none-any.whl Python 3 none any Details for spryng_analysis_mcp-0.2.0-py3-none-any.whl

Total release size: 172.7 kB

Release files / spryng_analysis_mcp-0.2.0.tar.gz

Download URL spryng_analysis_mcp-0.2.0.tar.gz
Size 80.5 kB
Tags Source
SHA-256 checksum
How to use checksums
2e4d9df1877a3b86c04dbea9a94e8e9e4eb725ad64f85c80f5d8adca4c85a82d
BLAKE2b-256 checksum
How to use checksums
3187945e440ef61743ba5fa8f519ff6fb48bf56eddef42547c783f91e115ba35
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Release files / spryng_analysis_mcp-0.2.0-py3-none-any.whl

Download URL spryng_analysis_mcp-0.2.0-py3-none-any.whl
Size 92.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
798b1d2628ea912351503e007c19179c65eafc43ae900cd185e020c585c8fe4b
BLAKE2b-256 checksum
How to use checksums
c7062aa0f02fa1cf321980f8f3a5c895c166034cf29c0bf43905e7994a899a59
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Security for spryng-analysis-mcp 0.2.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page