Pinlo
Use cases/Trading agentsForecasting agentsAsset researchField robots

Due diligence that accumulates

Memory for asset research.

Tokenized stocks, funds and bonds come with documents that change: terms, notices, corporate actions. Pinlo lets a research agent keep each version and each update it reviewed, and prove which ones it relied on.

Example

Research agent

Request

The issuer posted a notice about a share split. What changes for the tokens we track?

Recalled from the mind3 records

  1. 01Terms v3Token tracks one share; dividends adjust a balance multiplier, not the raw balance.
  2. 02ContractAddress and decimals read on-chain on 9 Oct; matches the issuer's published list.
  3. 03Prior noteLast split was applied through the multiplier within one trading day.
  1. 01Read the notice
  2. 02Recall terms and contract
  3. 03Write the impact note

Illustrative request and records, made up to show where memory fits.

The problem

What slips away
between sessions.

01

Versions pile up

Terms documents get revised. Without memory, an agent cannot tell which version it summarised last month.

02

Look-alikes confuse lookups

Many tokens share a ticker. An agent that re-checks from scratch each session can pick a different contract each time.

03

Updates arrive one by one

Splits, dividends and notices land weeks apart, and each only makes sense next to the ones before it.

With Pinlo

What this agent's memory holds.

01 / Document versions

Save each version you read

Store a summary of each terms document with its date and source. The hash fixes the exact text the agent summarised.

TermsVersionSource

02 / Checked identity

Remember the right contract

Save the contract address and the checks it passed as a fact record, so later sessions reuse a verified answer.

AddressDecimalsChecked on

03 / Event history

Corporate actions in sequence

Each notice becomes a record tagged with the asset. Recall by tag returns the full history in the order it was saved.

SplitDividendNotice

04 / Team access

One research mind, many readers

Grant read access to the agents that answer users, while only the research agent writes.

WriterReadersExpiry

The workflow

How the work
uses memory.

  1. 01

    Capture the update

    The agent saves the notice with its source link and the asset tag.

  2. 02

    Recall the asset file

    It recalls the current terms, the verified contract and earlier corporate actions before drawing a conclusion.

  3. 03

    Save the impact note

    Its summary of what changes is saved as a decision record and included in the next anchored version.

Where the line is

Pinlo does not verify issuers or give investment advice. It keeps what your agent read and concluded, and proves that record stayed the same after anchoring. Whether the source was right is a separate question.

Setting it up

Shape the records.
Connect the agent.

Quick start
  • Open a mind per asset, labelled with the ticker and the verified contract address.
  • Save identity checks (address, decimals, issuer list) as the first fact records.
  • Tag each update by kind: terms, dividend, split, notice.
  • Anchor after every document version you summarise.

Questions

Worth asking
first.

Does Pinlo check that a token is genuine?

No. Your agent does the checking; Pinlo remembers the result and when it was made, so the check is not repeated or lost.

Are the documents stored in full?

You decide. Most teams store summaries and links. Each record holds up to 2,000 characters of text.

Who can read the research?

The mind's owner and any wallet it granted read access to. Records on the hosted store are plain text unless you encrypt them in the browser first.

Is this investment advice?

No. Pinlo is memory infrastructure. It does not rate assets or recommend trades.

Related use cases

Let your research agent build a file, not a fresh start.