Skip to main content
SMEme
Create account Log in

You demand AI automation for efficiency — but you need AI automation you can control.

You have already surrendered control to an LLM if you're expecting it to interpret rules and context from prompts and saved runbooks.

SMEme helps you encode your rules into user-friendly decision trees you approve. SMEme converts those trees into decision constraints and provides logical analysis with auditable reasoning traces that document evidence of human oversight and explainability, something no LLM can do. SMEme works directly in chat to add the missing reasoning layer.

Just one MCP connection. No new software, knowledge engineers, consultants, or specialized training. Try SMEme now!

Go to the quick start instructions and copy prompts to get your chatbot to connect, help you identify candidate decision procedures, build your first decision tree, and run decision support analysis.

How it works

Codify the Decision Tree

You define decision constraints as simple decision trees your chatbot helps you build using your own policies, saved prompts, and AI assisted web research.

Video placeholder

Codify Your Judgment

Researched build

Get Decision Support

SMEme converts those decision trees into logical formulas for advanced decision support and logical analysis. All easily accessible in natural language directly in chat and programmatically for coded workflows.

Video placeholder

Consult Your Judgment

A fictional client folder

LLMs are great at interpreting context but you need provably correct reasoning. You need Rational AI.

A single MCP connection provides a rational, System 2, logical reasoning tool with no specialized training, software project, or lock-in.

  • Precision. Answers you can stand behind: the same facts produce the same conclusion every time, with a traceable decision path rather than an AI judgment call.
  • Built-in explainability. Ask which facts drove a conclusion, run what-ifs, or see what would have to change to reach a different outcome.
  • Expertise captured and scaled. Build decision procedures in plain chat from your own policies, documents, and research.
  • No new project. Avoid hiring knowledge engineers, building ontologies, or new software project budgets.
  • Works where you already work. A single connection plugs into the AI tools and automated workflows you already use.
  • No lock-in. You own your decision logic and can download, share, or delete it anytime, or self-host the engine in your own environment.
  • Humans stay in charge. Nothing goes live until an expert reviews and approves it, and the reasoning runs outside the AI model.

Rational AI means trust and control over the answers you receive and the automation they drive.

SMEme doesn't just apply rules or walk a flowchart. It reasons logically over the constraints you define in the decision tree. Ask it questions no chatbot can prove.

  • What follows

    A conclusion follows, is ruled out, or still depends on facts you have not admitted.

    “Can we accept this client?”

  • Which facts decided this

    The smallest set of admitted facts that still forces the conclusion.

    “Which facts made this a decline?”

  • What if a fact changed

    Compare this case with one different fact and see whether the conclusion still follows.

    “Would we still decline if the owner were domestic?”

  • What would have to change

    The edits that would make another conclusion follow, or leave it still possible.

    “What would it take to accept them?”

  • What cannot change the result

    Unanswered questions that would not move a conclusion that already follows.

    “Does the screening report still matter?”

  • What the tree allows

    The conclusions the deployed tree can reach.

    “What outcomes can this tree reach?”

Use cases

How will you SMEme?

Client work

  • Gmail
  • Google Drive
  • Salesforce
  • SMEme

Accept a new client

“Run my client acceptance decision tree on this prospective client. Read the Gmail thread, the intake sheet in Google Drive, and the Salesforce matter record.”
  • Salesforce
  • Teams
  • Outlook
  • SMEme

Catch a conflict

“Run my conflict check decision tree on this name. Look through the Salesforce matter list, the Teams channel, and recent Outlook mail.”
  • Google Calendar
  • Gmail
  • Google Chat
  • SMEme

Prep the intake call

“Before today’s call, run my client acceptance decision tree. Gather the Google Calendar invite, the Gmail thread, and the notes in Google Chat.”
  • Outlook
  • OneDrive
  • SMEme

Recheck a new fact

“A new Outlook message says the ownership changed. Run my client acceptance decision tree again with it and the chart in OneDrive.”

Engineering and Security

  • GitHub
  • Slack
  • SMEme

Ship the change

“Run my release decision tree on this pull request. Read the diff, the linked issues, and what we said in Slack.”
  • Notion
  • Atlassian
  • SMEme

Check the spec

“Run my scope decision tree on this request. Read the spec and the related project work.”
  • Linear
  • Gmail
  • Google Chat
  • SMEme

Triage the issue

“Run my triage decision tree on this issue. Read the issue, the customer Gmail, and the Google Chat thread.”
  • Sentinel
  • Teams
  • SMEme

Escalate the alert

“Run my alert triage decision tree on this incident. Pull the alerts and sign in history from Sentinel, and check the Teams channel for a change that explains it.”

Licensed professionals

  • Salesforce
  • OneDrive
  • Outlook
  • SMEme

Decide the withholding

“Run my cross border tax withholding decision tree on the ACME matter, the sale of asset X to party D. It is in Salesforce as matter 123. Read the project files in OneDrive, the intake memo, and the Outlook mail.”
  • athenahealth
  • Google Drive
  • Gmail
  • SMEme

Check medical necessity

“Run my medical necessity decision tree on this order. Read the chart in athenahealth, the clinical note in Google Drive, and the payer policy in Gmail.”
  • Salesforce
  • Google Drive
  • Gmail
  • SMEme

Review the recommendation

“Run my suitability decision tree on this recommendation. Read the Salesforce client profile, the product disclosure in Google Drive, and the compliance note in Gmail.”
Go to the quick start instructions and copy prompts to get your chatbot to connect, help you identify candidate decision procedures, build your first decision tree, and run decision support analysis.

For builders

Call SMEme programmatically from workflows.

Add more control to your automation by programming your own harness: evidence collection, human-in-the-loop interrupts, and downstream processing of SMEme results.

  1. Setup bootstrap Pick the tree Choose a Listed decision tree and open a session.
  2. Probabilistic propose Model proposes One of the offered answers, with its source.
  3. Human admit LangGraph interrupt A person admits, edits, or rejects. Rejections never reach SMEme.
  4. Deterministic evaluate SMEme evaluates The deployed tree decides over MCP.
  5. Report Route on result_kind to your next node.
    Isolated evaluation required SMEme stops instead of faking verification.
  6. ↺ Next question goes back to propose

Here is one LangGraph example of that harness. The code and how to run it are in the LangGraph docs.

Resources

DocsMCP docsConnect the chatbot you already use. WritingArista Labs blogEssays on judgment, expertise, and agency. SourceSMEme Core on GitHubSource available under the SMEme SUL 1.0.

Questions

What is SMEme?

SMEme connects to the chatbot you already use. You authorize a decision tree. SMEme compiles that tree and a constraint solver determines what follows.

What is a decision tree?

The procedure you can read. Questions, branches, and outcomes, in plain language. You see the tree. SMEme compiles the authorized tree into an intermediate representation and reasons over it as boolean constraints.

New client
ConflictDecline
CapacityAccept
Ownership unclearIt depends
Do I need to be technical?

You don’t have to know how to explain it. You only have to recognize when a draft is right, wrong, or incomplete. No coding, no special syntax, and no knowledge engineers.

What are constraints and evidence?

Constraints are the rules a decision must satisfy, such as your policies, the law, and your own judgment. After you deploy the tree, SMEme applies those rules as written. Evidence is context that has been used to answer a question in your decision tree, with its source attached. Your data stays in the apps you already use.

Why not just ask the chatbot?

A chatbot goes from context to an answer. SMEme puts evidence in the middle and constraints in charge. The reasoning follows the tree you deployed. The facts stay visible. If the evidence conflicts or is incomplete, the answer stays unresolved.

What happens after I build a tree?

You authorize in the editor by choosing Deploy. That saves the version SMEme will reason over. Then set it to Listed, which means it shows up among the SMEme tools in your chat.

Where does my data go?

Your files stay in the apps you already use. The chatbot turns context from those apps into evidence for a question in the tree. The tree does not keep the matter files.

Can I share a tree?

You can download your decision tree. It stays on your account. Giving another account its own copy is not available yet.

What does it cost?

Hosted Free includes 3 decision trees, 150 MCP calls, and 3 SMEme AI tree builds. Deleting a tree frees a slot. It does not add calls. See pricing for Pro and self hosting.

Can developers build with SMEme?

SMEme Core is source available under the fair code SMEme SUL 1.0. The hosted site is a separate service. Start with the MCP docs and the Core repository.

Our mission is to make AI increase human agency.

AI should expand judgment you can stand behind, not bury consequential decisions in model weights. You write and authorize the decision tree. SMEme reasons over it. Every tree is yours to download, share, and delete. Read our Sovereignty and AI Principles and Privacy Policy.