SentiSift SDKs

SentiSift SDKs

Official client libraries for the SentiSift comment-moderation and intelligence API. Three published packages, all open source for the integration code so developers can read, audit, and contribute.

What is SentiSift?

SentiSift filters bots, spam, and noise out of comment sections without silencing real voices. It scores each comment on five independent axes (sentiment, eloquence, length, behavioral patterns, commercial signals) and returns moderation decisions plus crowd-level analytics. On paid tiers, it also contributes constructive perspectives when a discussion skews one-sided ("Influence").

Get a free API key (1,000 comments, no credit card) at sentisift.com/pricing. Full API reference: sentisift.com/api-docs.html.

Quick start

Python

pip install sentisift

from sentisift import SentiSift

client = SentiSift()  # reads SENTISIFT_API_KEY from env
result = client.analyze(
    article_url="https://example.com/article/1",
    comments=[
        {"text": "Great article!", "author": "alice", "time": "2026-04-18T10:00:00"},
    ],
)

Node / TypeScript

npm install @sentisift/client

import { SentiSift } from "@sentisift/client";

const client = new SentiSift();
const result = await client.analyze({
  articleUrl: "https://example.com/article/1",
  comments: [{ text: "Great article!", author: "alice", time: "2026-04-18T10:00:00" }],
});

MCP (Claude Desktop, Cursor, VS Code, Continue)

uvx sentisift-mcp     # one-shot run via uv
# or
pip install sentisift-mcp     # then add to your MCP host's config

Then add the SentiSift entry to your host's MCP config (Claude Desktop's claude_desktop_config.json, Cursor's ~/.cursor/mcp.json, VS Code's settings.json, etc.). See mcp/README.md for host-specific snippets.

Repository layout

.
├── python/                Python SDK (sentisift on PyPI)
├── node/                  Node SDK (@sentisift/client on npm)
├── mcp/                   MCP server (sentisift-mcp on PyPI)
├── OVERVIEW.md            Endpoint coverage matrix, supported platforms, release history
├── RELEASE_RUNBOOK.md     AI-driven release procedure (read this for any version bump)
├── RELEASE_CHECKLIST.md   Engineering line-by-line checklist
└── .gitlab-ci.yml         CI: tests on every push; publishes on tag push

Versioning

All three packages follow Semantic Versioning. Pre-1.0 minor versions may include breaking changes. Per-package CHANGELOG.md files document every release.

Contributing

Bug reports and feature requests welcome via issues. For pull requests, please open an issue first to discuss the change.

Architecture

The full SentiSift API service (auth, billing, scoring pipeline, scrapers) lives in a separate private repository. This public repository contains only the client-facing packages and their integration code, OpenAPI-derived models, and tests.

Documentation is hosted on sentisift.com:

• HTML reference: api-docs.html
• LLM-friendly Markdown mirror: api-docs.md
• OpenAPI 3.1 spec: openapi.json
• AI-agent integration guide: AGENTS.md
• llmstxt.org index: llms.txt

License

The SDK source code in this repository is MIT licensed — see LICENSE. The SentiSift API service itself is a separate hosted service governed by the Terms of Service at sentisift.com/terms.html. Use of the API requires an API key obtained from sentisift.com/pricing.

Contact

• General: [email protected]
• Customer support: [email protected]
• Issues: GitLab issues