Anti-hallucination Research Pipeline

Sanusi Mubaraq — Thu, 26 Mar 2026 21:55:10 +0000

This is a submission for the Notion MCP Challenge

What I Built

A CLI research pipeline that turns a free-form topic into structured, evidence-backed claims and optional narrative output, with Notion as the place those results live.

Research and fact-check. The tool expands the topic with an LLM (brief plus multiple web searches), merges and deduplicates what Tavily returns, then asks an LLM to pull out discrete claims tied to specific URLs and snippets. Each claim is challenged in a second pass: another search for counter-evidence and an LLM judgment with a confidence level, a contradiction flag, and short notes—so every assertion carries a provenance trail, not just a paragraph of prose.

Optional report. You can also generate a readable research write-up (Markdown) and publish the same content as a normal Notion page—separate from the structured claim data.

Notion sync. A companion step pushes the claim set into Notion as a Research claims database: one row per claim, with the claim text, source link, snippet, topic label, confidence, contradiction flag, and fact-check notes. Sync uses the Python MCP client against Notion’s hosted MCP.

Re-runs. You can refresh adversarial fact-checking on an existing claim set without redoing the whole research pass.

Stack: Python, OpenRouter (LLM), Tavily (search), and Notion MCP for the sync step. Useful for research, due diligence, or keeping a living backlog of traceable assertions in Notion.

Video Demo

Show us the code

https://github.com/LPMatrix/notion-mcp

How I Used Notion MCP

Notion MCP is the only way data gets into Notion in this project. The research pipeline — search, claim extraction, adversarial fact-check — writes to a local JSON file. A separate script, sync_to_notion.py, handles the sync entirely through MCP:

Connects to Notion's hosted MCP server at https://mcp.notion.com/mcp using the Python MCP SDK with Streamable HTTP transport and an OAuth Bearer token
Calls list_tools() after session.initialize() and resolves tool names at runtime
Creates the "Research claims" database via the create-data-source tool, passing the parent page ID, title, and property schema
Creates one page per claim via the create-page tool, writing all seven properties: Claim, Source URL, Source Snippet, Topic, Confidence, Contradiction, and Fact-check notes

Yield Farming

Sanusi Mubaraq — Mon, 31 Jan 2022 04:22:50 +0000

According to coinbase, DeFi (or “decentralized finance”) is an umbrella term for financial services on public blockchains, primarily Ethereum. With DeFi, you can do most of the things that banks support — earn interest, borrow, lend, buy insurance, trade derivatives, trade assets, and more — but it’s faster and doesn’t require paperwork or a third party. As with crypto generally, DeFi is global, peer-to-peer (meaning directly between two people, not routed through a centralized system), pseudonymous, and open to all.

Yield farming is an application of DeFi. This is lending or staking your cryptocurrency so you can get reward from it. This is somewhat almost like earning interest from a bank account;. Only yield farming are often riskier, volatile, and sophisticated unlike putting money in a bank.

How does Yield Farming Works

Users providing their cryptocurrencies for the functioning of the DeFi platform are known as liquidity providers (LPs). These LPs provide coins or tokens to a liquidity pool—a smart contract-based decentralised application (dApp) that contains all the funds. Once the LPs lock tokens into a liquidity fund they are awarded a fee or interest generated from the underlying DeFi platform the liquidity pool is on.Put simply, it is an income opportunity by lending your tokens through a decentralised application (dApp). The lending happens through smart contracts with no middleman or intermediator.The liquidity pool powers a marketplace where anyone can lend or borrow tokens. The usage of these marketplace incurs fees from the users, and the fees are used to pay liquidity providers for staking their own tokens in the pool.
The interest in an yield farm investment is calculated in two ways: APY - Annual Percentage Yield (interesting rate including compound interest) and APR - Annual Percentage Rate (interest rate exluding compound interest)

Yield Farming Platforms

There are various yield farm protocols used by yield farmers to stake and earn interest from their cryptocurrency.Listed below are some of these platforms.

Compound Finance

Compound is an open source protocol built for developers that uses an algorithmic, autonomous interest rate protocol to determine the rate depositors earn on staked coins.

Curve Finance

Curve is an exchange liquidity pool on Ethereum designed for: extremely efficient stablecoin trading, low risk, supplemental fee income for liquidity providers, without an opportunity cost.

AAVE

Aave is an Open Source Protocol to create Non-Custodial Liquidity Markets to earn interest on deposits and borrow assets with a variable or stable interest rate. The protocol is designed for easy integration into your products and services.

Uniswap

Uniswap is a decentralized exchange (DEX) where liquidity providers must stake both sides of the pool in a 50/50 ratio. In exchange, you earn a portion of the transaction fees plus UNI governance tokens.

Conclusion

Yield farming involves staking, or locking up your cryptocurrency in exchange for interest or more crypto. While it's possible to earn high returns with yield farming, it is also incredibly risky. A lot can happen while your cryptocurrency is locked up, as is evidenced by the many rapid price swings known to occur in the crypto markets.

DEV Community: Sanusi Mubaraq