SOP & Tribal-Knowledge Engine
a shop where the real how-to lives in one long-tenured employee’s head and a dozen half-written files nobody can find.
Drop a folder of your rough notes, Slack exports, and onboarding docs in, and it turns each one into a titled, numbered procedure and builds a search index so a new hire can type “how do I reset a customer’s password” and get the exact step back, not a pile of documents.
$39 a month. The cheapest kind of insurance against the day your best person is out sick and nobody else knows the closing routine. Cancel any day and the written procedures stay yours.
The knowledge that walks out the door at 5pm
Every small business runs on a handful of things only one person actually knows. How the register gets counted. Which supplier to call when the usual one is out. The order the ovens get turned on. It works fine until that person quits, goes on leave, or just isn’t reachable on a Saturday, and suddenly a routine task is a phone call and a guess.
Writing it all down is the obvious fix and nobody ever does it, because writing it up properly and then remembering where you filed it is its own boring job. The cost isn’t dramatic. It’s the slow tax of retraining, of mistakes made by people who were never told, of the same question answered out loud for the fourth time.
Walk, structure, index
- It walks your source folder recursively for
.txtand.mdfiles, skipping anything overMAX_FILE_MBso a giant transcript doesn’t stall it. - It builds each procedure by pulling a title from the first heading, splitting the body into ordered steps from your numbered lines, bullets, or plain sentences, and tagging it with the eight most significant words it found.
- It builds a search index as a plain term-frequency map, no model and no network call, so the same question against the same notes always returns the same ranked answer.
What lands in the output folder
- A clean SOP markdown file per source document, in a
SOPs/folder, readable by a human without any app. - A
SOP_LIBRARY.mdcatalog so you can see the whole set at a glance. - An
index.jsonthat powers the search, built once so lookups are instant. - A search command that answers a plain-English question by naming the best-matching procedures and quoting the single best step from each, which is what someone actually needs when they’re mid-task and stuck.
- Tags on every procedure, so a topic like “refunds” or “onboarding” pulls its related write-ups together without you organizing anything by hand.
Who it fits, and who it doesn’t
Teams of a few people up to a couple dozen who have the knowledge but not the documentation, and who don’t want to move their whole operation into a wiki platform to fix that.
Not for you if you already keep a tidy internal wiki that people genuinely search, or if what you actually want is a bot to invent brand-new procedures from scratch. This organizes and finds what you’ve written. It doesn’t compose policy you never wrote.
What it replaces, and what those do better
A shared Notion or Google Doc wiki, a binder in the back office, and mostly the senior employee you keep interrupting. A wiki tool is better at rich editing, images, and letting several people co-write a page in real time. The senior employee is far better at the judgment call, the weird edge case the notes never covered.
What those don’t do is turn the pile you already have into structured, searchable procedures in one run, offline, without anyone learning a new interface. This does the write-up-and-file chore so the knowledge stops living in one skull.
Where your notes stay
There’s no LLM and no cloud call anywhere in this. Search is deterministic ranking over a JSON index built on your own machine, and it runs with the network unplugged. Everything it reads stays inside your source folder, everything it writes stays inside your output folder, at 0600 permissions.
The procedures come out as markdown files you own. Stop paying and they’re still there, still readable in any text editor, because they never lived anywhere but your disk.
$39 a month, and getting the first library
Set the folder your rough notes live in and the folder where you want the procedures written. Run the build. You get a full library and a working search in a couple of minutes, and re-running it after you add new notes just refreshes the set.
Security and privacy
Reads a company’s internal notes/transcripts, which may contain sensitive operational detail. Security is designed in.
Principles
– Local-first. Nothing touches the network. No cloud dependency, no telemetry, no LLM API calls, no external calls of any kind. Search is deterministic term-frequency ranking over a locally built JSON index.
– Read-only against the source. The tool never writes, moves, renames, or deletes anything under SOP_SOURCE_FOLDER. It only reads .txt/.md files there and writes its own generated library under OUTPUT_FOLDER.
– Least privilege. Reads SOP_SOURCE_FOLDER (build only), writes only under OUTPUT_FOLDER.
Protections implemented (v0.1.0)
– Path-traversal defense — every generated file path (SOP markdown, index.json, SOP_LIBRARY.md) is checked with assertWithinRoot() before writing, and derived filenames are sanitized with safeSegment(), so a crafted source filename can’t escape OUTPUT_FOLDER.
– Fail-fast config validation — clear errors for missing SOP_SOURCE_FOLDER (build mode), missing OUTPUT_FOLDER, or invalid MAX_FILE_MB/TAG_COUNT before any file is touched.
– Restrictive permissions — generated files (SOPs/*.md, index.json, SOP_LIBRARY.md) are written 0600.
– Secrets discipline — .env and the generated SOP library output are git-ignored; no secrets or company data in the repo.
– Bounded reads — files larger than MAX_FILE_MB are skipped entirely (logged in the catalog as skipped) to bound memory and runtime on large note dumps.
– File-type scoping — only .txt and .md files are read from SOP_SOURCE_FOLDER; everything else is ignored during the walk.
Known limitations / roadmap
– Search ranking is term-frequency (TF) over an inverted index — it has no synonym awareness, stemming, or semantic understanding. A query using different words than the source note (“password reset” vs. “credential reissue”) may not match well; this is an intentional trade-off for determinism and zero external dependency.
– Title/step extraction is heuristic (headings, numbered lines, bullets, sentence splitting) and works best on notes with some existing structure; unstructured stream-of-consciousness transcripts will produce a single-sentence-per-step SOP that likely needs manual cleanup.
– No file-content type verification beyond extension filtering (.txt/.md) — the tool trusts file contents to be plain text; it does not attempt to open other formats.
– Rebuilding overwrites SOPs/index.json/catalog in place; there is no versioning or diffing of prior builds yet.
Reporting
Report suspected vulnerabilities privately to the maintainer before public disclosure.
Frequently asked questions about SOP & Tribal-Knowledge Engine
What is the SOP & Tribal-Knowledge Engine and how does it work?
The SOP & Tribal-Knowledge Engine takes the know-how that usually lives in one person’s head and turns it into something the rest of your team can actually ask. In most small businesses the real instructions are scattered. Part is in an old document, part is in a manager’s memory, and part is nowhere except the way things have always been done. When that person is out sick or leaves, the business slows to a crawl. This tool is built to close that gap by capturing what you already know and making it answerable on demand.
The setup is plain. You upload or record your existing standard operating procedures and process notes. The engine indexes them into a searchable knowledge base. Then your staff ask questions in their own words and get answers pulled straight from your procedures, not from generic advice off the internet. If a new hire asks how you close out the register or handle a return, the answer comes back grounded in your own written steps. That is the difference between a pile of documents nobody reads and a system that hands the right step to the right person at the moment they need it.
Many of those procedures quietly touch your records. How you file receipts, how long you keep invoices, how you archive a completed job all belong in an SOP, and they connect to real duties. the IRS guidance on how long to keep business records sets out how long a business must hold onto the documents behind its return, and writing that retention rule into your procedures is how it actually gets followed. When the rule lives only in the owner’s head, it gets skipped the first busy week. When it lives in the knowledge engine, anyone can check it in seconds.
The tool is aimed at the owner whose processes have outgrown memory and sticky notes. Instead of retraining each person from scratch, you build the knowledge once and let staff draw on it whenever they are stuck. You can see how the capture-and-ask loop runs on the product page, which walks through a first setup. It is not a course-builder that makes you author polished lessons, and it is not a wiki you have to police. It meets you where your knowledge already is, messy and half-recorded, and makes it useful.
The name points at the real target. Tribal knowledge is the unwritten stuff a team just knows, and it is invisible until the person carrying it is gone. A retirement, a resignation, or a single bad flu week can take years of hard-won routine with it. Writing that knowledge down after the fact is nearly impossible, because no one else remembers the details. Capturing it while the person is still around is the only reliable fix, which the tool makes painless.
The honest promise is narrow. The SOP & Tribal-Knowledge Engine does not invent procedures you never wrote, and it will not run your business for you. It answers from what you feed it, which means the quality of its answers tracks the quality of your inputs. Feed it your real steps and it becomes the memory your team can rely on when the one person who knew everything is unreachable. For small-business operators that alone is often the difference between a smooth shift and a stalled one.
How does the engine help me document payroll and timekeeping procedures?
Payroll and timekeeping are exactly the kind of process that hides in one person’s routine until that person is away. How hours get recorded, how overtime is approved, how a pay period gets closed, and where the records end up are all steps that a business runs on habit far more than on paper. The SOP & Tribal-Knowledge Engine is built to pull those habits into writing and keep them answerable, so the process survives a vacation, a resignation, or a busy month.
You record or upload how your business handles time and pay, and the engine indexes it into a knowledge base your staff can query. A manager covering payroll for the first time can ask how a timesheet gets approved and get your actual answer, not a guess. That matters because the steps are not just internal preference. DOL Fact Sheet 21 on FLSA recordkeeping explains that employers must keep payroll records and wage-computation records for set periods, two and three years depending on the record. If the person who knew where those files lived is gone, that duty gets a lot harder to meet.
By capturing the procedure once, the tool makes the recordkeeping repeatable instead of personal. Where you store timecards, how long you retain them, and who signs off become written steps anyone can follow. That is the practical value of documenting a process rather than trusting memory. When an inquiry or an audit asks how you handle wage records, you can point to a procedure that people actually followed, because it was sitting in the knowledge engine the whole time rather than locked in someone’s head.
The engine does not calculate payroll and it is not a timekeeping clock. It will not replace whatever runs your wages. What it does is hold the instructions for using those tools correctly and consistently, so the knowledge does not walk out the door with an employee. That is a deliberately narrow job. The value is in continuity, making sure the fifth person to run payroll does it the same careful way as the first.
Picture the week your bookkeeper is out. Timecards still need approving and the pay period still has to close. Without a written process, whoever steps in is guessing, and a guess on payroll is how a business ends up with wrong paychecks or missing records. With the procedure captured, the stand-in asks the engine and follows the same steps your bookkeeper would. The work gets done correctly, and the records land where they belong, which is the whole reason to write it down before you need it.
For an owner, this turns a fragile, memory-based process into a documented one without a heavy rollout. You do not have to write a formal manual in one sitting. You capture the steps as they come up and let the SOP & Tribal-Knowledge Engine organize them. You can read how that incremental capture works on the product page. The result is a payroll and timekeeping process that stays intact through turnover, which is precisely the kind of resilience that keeps small-business operators out of trouble when the one person who knew it all is suddenly unavailable.
Where do my procedures and staff questions live, and is that data safe?
Any tool that captures how your business really runs is holding something valuable, so it is fair to ask where that content lives and who can reach it. The SOP & Tribal-Knowledge Engine works from the procedures you upload or record and the questions your staff ask against them. That means the sensitive material is your operational know-how, and sometimes the names or details that show up inside a procedure. Knowing what the tool holds is the first step to handling it responsibly.
The plain-language standard worth applying comes from the FTC guide Protecting Personal Information, which tells businesses to take stock of the data they hold, scale down to what they truly need, lock down what they keep, and pitch what they no longer use. That checklist fits a knowledge tool well. Take stock by knowing which procedures contain sensitive detail. Scale down by not pasting customer records into an SOP that does not need them. A process document should teach the step, not become a second copy of your customer list.
Access is the other half. A knowledge base is only as safe as the login in front of it. Staff should reach the engine through real accounts rather than a shared password taped to a monitor, so you know who asked what and can cut off access when someone leaves. Because the tool answers from your own procedures and keeps them in one place, you have a single spot to review and control rather than the same knowledge smeared across email threads, chat histories, and personal notebooks nobody can audit.
It also helps to keep procedures and personal data separate on purpose. The point of the SOP & Tribal-Knowledge Engine is to capture how work gets done, not to warehouse the private details of customers or staff. Kept narrow, it lowers your exposure rather than adding to it, because there is less sensitive information sitting inside it in the first place. That is the same instinct the FTC guide points every small business toward, scale down before you worry about locking down.
Disposal deserves a mention too. Procedures go stale as the business changes, and an outdated SOP can mislead as badly as no SOP at all. Reviewing and clearing old material keeps the knowledge base trustworthy and shrinks what you are holding, which is the pitch-it step the FTC guide describes. A tool that centralizes your procedures makes that housekeeping a single, visible task rather than a hunt across scattered files you have long forgotten exist.
The honest guidance is that no software removes your responsibility to handle information with care. What this tool does is make careful handling easier by centralizing your procedures under real access control and giving you one place to review what is stored. You can see how the knowledge base is set up on the product page. For small-business operators the takeaway is simple. Capture the know-how, keep the private data out of it where you can, and put a real login in front of the rest. Safe handling is a habit, and a single well-controlled knowledge base makes that habit far easier to keep than a dozen scattered documents ever could.
How is the SOP & Tribal-Knowledge Engine different from Trainual, Scribe, or Notion?
Tools like Trainual, Scribe, Notion, and Process Street are good at one thing, storing documents. You write a procedure, save it, and hope someone opens it later. The problem is that people rarely go hunting through a document library in the middle of a task. They ask the nearest coworker instead, which is exactly how knowledge stays trapped in one person’s head. The SOP & Tribal-Knowledge Engine is built around the asking, not just the storing, and that is the core difference.
Instead of leaving your procedures as files to browse, the engine indexes them into a knowledge base and lets staff ask questions in plain language, then answers from your own recorded steps. A new hire does not need to know which document holds the answer or what it is titled. They ask, and the tool retrieves the relevant procedure and responds with it. That turns a passive archive into something closer to the experienced coworker who always knew where everything was, except it never leaves and never gets annoyed at the tenth question.
There is a useful way to frame why this matters beyond convenience. the NIST Cybersecurity Framework organizes reliability work around identifying your important assets and protecting them. Your operating knowledge is one of those assets, and when it lives only in memory or a rarely opened wiki, it is neither identified nor protected. Capturing it in the tool makes it an asset you can actually see, maintain, and hand to the next person. That is a different goal from a document store, which identifies nothing and simply holds whatever you dropped in.
The practical upshot for an owner is that you are not choosing a prettier place to park documents. You are changing how the knowledge gets used day to day. A stored procedure helps only the person disciplined enough to search for it. An answerable procedure helps everyone, including the people who would never have looked. You can see how the ask-and-answer flow compares on the product page, which lays out the setup for a small team.
There is also a maintenance difference. A document store grows heavier over time as files pile up and go stale, and no one is responsible for weeding it. Because the engine is built around answering, keeping it accurate has a clear payoff in better answers, so there is a real reason to prune and correct. That feedback loop is missing from a passive archive, where a wrong document sits undisturbed until it burns someone who trusted it.
The blunt truth is that most SOP tools fail not because the documents are bad but because nobody reads them. The SOP & Tribal-Knowledge Engine is aimed at that failure directly. It does not try to out-format a wiki or out-design a note app. It makes your existing procedures respond when someone needs them, which is the step every document-first tool skips. For small-business operators whose knowledge is stuck in one head, that shift from storing to answering is the whole reason to bother.
Can staff really get straight answers from our own recorded procedures?
Yes, and that is the heart of what the SOP & Tribal-Knowledge Engine does. The value is not a fancy search box. It is that a staff member can ask a real question, in their own words, and get an answer grounded in the procedure you actually wrote, rather than a generic tip that may not match how your business works. You record or upload your steps, the engine indexes them, and from then on the answers come from your material and nowhere else.
Think about the everyday moments this covers. A cashier is unsure how to process a specific refund. A weekend lead cannot remember the closing checklist. A new hire needs to know who approves a discount. In each case the honest need is the same, a fast answer that matches your rules. The tool retrieves the right procedure and hands it over, which keeps the work consistent no matter who is on shift. Consistency is what protects a small business from the slow drift where every person does the same task a slightly different way.
Because staff query the knowledge base through the engine, you also get a natural place to manage who can see what. That access matters, because your procedures can touch sensitive parts of the operation. the FTC Safeguards Rule describes the kind of written program a business should keep for protecting customer information, built on administrative, technical, and physical safeguards. Putting your operating knowledge behind real accounts rather than a shared login is a small, sensible piece of that same discipline, and the tool makes it the default rather than an afterthought.
There are honest limits worth stating. The engine answers from what you give it. If a procedure is missing, vague, or out of date, the answer will reflect that, so the tool works best when you keep feeding it and correcting it as the business changes. It is a memory for your team, not an oracle. It cannot tell staff how to do something you never wrote down, and it will not quietly make up a policy to fill a gap.
It is worth stressing that the answers are only as good as the procedures behind them, which is why the SOP & Tribal-Knowledge Engine is built to keep improving as you use it. Every question a staff member asks is a signal about what your procedures cover well and where they fall short. Over time you close those gaps, and the engine gets sharper at handing back the right step. That living quality is what separates it from a manual you wrote once and never touched again.
For an owner, the payoff is fewer interruptions and steadier work. The questions that used to land on the one person who knew everything now land on the tool, and the answers stay consistent whether that person is in the building or on vacation. You can read how the answering works on the product page. For small-business operators who have felt a whole shift stall because the right person was out, being able to pull a straight answer from your own procedures is exactly the kind of resilience that keeps the day moving.
| Attributes | Value |
|---|---|
| Deployment |
Cloud |
| Platform |
Web |
| License Type |
Subscription |
| Billing Period |
Monthly |
| Target Customer |
Small Business & Retail |
| Free Trial |
Yes |
Frequently bought together
-
Browser Workspace / Tab Context Organizer
$35.00 -
Data-Entry Killer (email -> records)
$35.00 -
Excel-to-Vertical-App Converter
$39.00 -
KPI Dashboard Autobuilder / Desktop Command Center
$39.00 -
Local Demand Forecaster
$39.00 -
Local Video Transcription / Private Note Taker
$75.00 -
Rental Property Micro-ERP
$24.00 -
SMB Ops Reconciliation OS / Owner Control Center
$75.00








