Sovereign expert reasoning — decision-trees your agents call through MCP
Not every task should run inside a language model — especially expert reasoning. SMEme helps you encode judgment as decision-trees, then compiles them into a logical analysis engine. Your agent gathers evidence; SMEme evaluates against your rules and returns a deterministic, attributable result. Language models handle language. SMEme handles analysis. On SMEme-hosted, deterministic evaluation runs on our infrastructure — not inside the LLM provider. Self-host with the source-available SMEme Core image when residency or policy requires your own environment.
neuro-symbolic AI at scale
SMEme encodes knowledge — not prompts
You see your decision tree as a decision tree; SMEme sees it as a micro-theory. Run full-range logical analysis from the LLM chat harness with real data it collects or with one-off hypotheticals. Results are returned direct to the agent as context to continue chatting or serve as inputs to runbooks/other agents. Rely on verifiable and transparent logical analysis of knowledge you author and control.
LLM-assisted expert system
Agent gathers facts → SMEme tests Reachability of conclusions
Your agent is connected to your data. Ask it to gather the relevant facts and call SMEme to evaluate what conclusions can be reached. Don’t dump the flow-chart into a saved prompt and hope for the best. Get verified answers from auditable decision trees while saving tokens.
Logical analysis assistant
Hypotheticals and how-to questions → SMEme returns verified reasoning traces
Ask natural language questions like, “what would have to change for me to get to [another specified conclusion]?” or, “what would happen if [fact x] were to turn out to be false?” Rich results are returned in a structured format directly into chat as context.
Start free. Scale as you grow.
Shared access (Business tier, Coming Soon) delivers conclusions only—your decision tree logic stays on SMEme, even when you share privately with clients.
Free Forever
Try SMEme end-to-end
$0 / month
No credit card required.
- 3 decision trees
- 150 MCP tool calls* / month
- 3 AI-assisted decision-tree builds / month
- MCP server access
Pro
For individual experts
$49 / month
Billed monthly. Cancel anytime.
- Up to 20 decision trees
- 1,500 MCP tool calls* / month
- 20 AI-assisted decision-tree builds / month
- MCP server access
- Email support
Business
Distribute to clients & colleagues
Pricing at launch
Join the waitlist for early access and launch pricing.
- Everything in Pro, plus private sharing & client monetization
- Higher limits than Pro — decision trees, MCP calls, and AI builds
- Private decision tree sharing — invite clients & colleagues
- Monetize decision-tree access with Stripe payouts
- Priority support
* How usage is counted
* MCP tool calls are weighted by operation type. Evidence validation and a standard decision tree evaluation each count as 1 (a typical agent session uses both); “what if” follow-ups count as 2; “how to” follow-ups (reaching a different outcome) count as more. Listing decision trees and downloading worksheets do not count toward your allowance.