70+ definitions across mesh, flagships, AI/LLM, compliance, and buyer terms. Last updated 2026-05-03.
This glossary is the canonical reference for every concept Manera uses across its product, contracts, marketing, and trust doctrine. Each term is a stable URL fragment (e.g. #mesh-tier) so you can deep-link from contracts, RFPs, or AI-engine citations.
Manera's flagship subscription tier at $999/mo (or $9,990/yr at the 15% annual discount). Mesh Tier unlocks every Manera Intelligence sub-app at the buyer-visible quality bar (28+/40 v2 audit floor) plus cross-app mesh queries. Designed for the single-approver economics threshold — under $1K/mo means no procurement committee, no RFP, no security review delay.
Manera's enterprise tier at $9,999+/mo (annual contract). Sovereign Tier adds dedicated infrastructure, customer-managed encryption keys, on-prem or VPC deployment, contractual SLA, and white-glove onboarding. For buyers who can't store regulated data on shared SaaS infrastructure (Loi 25, GDPR Art. 28, FFIEC).
A query that crosses two or more Manera sub-apps in a single answer — e.g. "what's our exposure to a sanctions cascade affecting suppliers in Country X?" hits NEIP (sanctions feed) + ProductDNA (supplier graph) + LexiWorld (regulatory clock) and returns one synthesized answer. Mesh queries are the structural moat — no single-product competitor can answer them.
The complete lineage of a Manera answer: which sub-apps were queried, which sources fed those sub-apps, what model produced the synthesis, when, with what cache-key. Every Manera answer carries a provenance ID surfaced via "Show provenance" — designed for EEOC, GDPR, and EU AI Act audits.
Append-only, tamper-evident log of every material decision made by a Manera flagship — each entry includes the previous entry's SHA-256 hash, so any modification is detectable. Used by Oracle (trade decisions), TalentIntel (hiring decisions), LexiWorld (legal advice), AdversarialAI (red-team findings).
Manera's structural commitment that every flagship must hold ≥ 80% gross margin against Claude API + infrastructure cost. Enforced by the shared/usage_governor.py 7-tier matrix — when a sub-app's margin drops below 80%, governance throttles new mesh queries until prompt-caching coverage or model routing recovers the gap.
The pricing principle that any Manera SKU under $1,000/mo can be approved by a single decision-maker (CFO, CISO, GC) without procurement committee or security review delays. Mesh Tier ($999/mo) is calibrated exactly to this threshold — one notch under the typical $1K corporate-card limit.
The Manera SDK pattern that wraps every Claude API call with pre-flight budget check + cache-control + provenance stamp. Internally calls usage_governor.pre_flight() before anthropic.messages.create(), raising BudgetExceededException if the call would breach the 80% margin floor.
The exception raised by usage_governor.pre_flight() when a Claude API call would breach the configured tier budget. Sub-apps catch this and surface a graceful "budget exceeded — try again next billing cycle or upgrade tier" response rather than silently 500-ing.
The bidirectional file-based partnership protocol between Manera's developer-builder agent (Metta) and the SATORI trading agent. Metta writes METTA_ADVISORY.md with risk rules; SATORI reads it on every chat startup and writes AUTO_CONSULT_LOG.md when it makes material decisions. Pre-trade check blocks any trade > $5K that would violate the advisory.
Manera's pricing doctrine — too cheap = asceticism trap (lose money + reputation); too expensive = indulgence trap (become Bloomberg, kill the swarm). Mesh Tier $999/mo is the calibrated balance: covers Claude COGS at 85% margin and stays under the single-approver threshold. Reviewed quarterly via the 6 defensive practices.
The minimum v2 audit score (8 axes, 40 points) a Manera sub-app must hold to be visible in Mesh Tier. Apps below 28 go preview/admin-only until lifted. As of May 3 2026 all 21 surviving Intel apps clear this floor — 100% Mesh Tier eligibility.
The percentage of a sub-app's Claude API token spend that benefits from prompt caching's ephemeral cache-control directive. Manera's portfolio average is 78% (Wave 9 May 3); the goal is 90%+ before Phase 2 capture. Each 10-point increase saves ~$10-15K/yr at $1M ARR.
The 2-petal Legal flagship — LexiWorld + IntelHub — at $999/mo. EUR-Lex + CanLII + GOV.UK case-law engine with breach-clock; Westlaw-class research at 1/30th the cost. Mesh integration with Strategy (regulatory cascade) and Cyber (breach-clock).
The 2-petal Real Estate flagship — RealEstatePulse + WeatherPulse — at $999/mo. Macro-housing strategist + REIT comps + cap-rate spread + climate exposure model. Built for CROs (Chief Real Estate Officers) at REITs and family offices.
The cyber War Room flagship — real-time AI-zero-day defense + outcome prediction + SHA-256 verifiable lineage. 38/40 v2 audit. Standalone $199/mo or included in Mesh Tier. The only flagship with a public live-status URL: nexusai.maneratech.com.
The Oracle + SATORI trading flagship — currently dormant for compliance review. SATORI is the multi-tenant trading agent (Track A shipped May 2026: auth + per-user isolation); Oracle is the SHA-256 audit-chain trade lineage layer. Returns to GA when ≥ 50 SATORI trades + Oracle /performance live ≥ 14 days.
FX hedging signal + execution sub-app inside Treasury. 22-currency coverage via Wise. Backtesting v2 with walk-forward + Monte Carlo + PDF report. Wise Personal API token gives read-only balances; OAuth unlocks one-click execution.
Bond + credit market intelligence sub-app inside Treasury. FRED + Treasury auctions + corporate spread monitor. Counterparty risk feed for new customer onboarding via mesh-link to MANERAbilling.
Narrative sentiment analysis sub-app — measures how the financial press is framing a topic, ticker, or event. Cross-app feed into NEIP (geopolitical narrative), Treasury (FX vol triggers), Strategy (M&A read-through).
Cross-flagship mesh orchestrator — the coordination layer that lets Strategy + Legal + Treasury answer a single mesh query. Sub-app inside Legal flagship by default; available across the mesh.
Threat intel sub-app inside Cyber. CISA KEV feed + 40+ threat-feed integrations. Mesh-link to ResiliencePulse for incident orchestration and to AdversarialAI for live red-team validation.
Endpoint detection + response (EDR) intel sub-app inside Cyber. Lighter than CrowdStrike Falcon, designed to surface signals you can act on rather than rebuild your SOC.
Phishing simulation + awareness sub-app inside Cyber. KnowBe4 alternative at 1/40th the price. Currently below the 22+/30 quality bar (re-positioning in progress).
AI red-team sub-app inside Cyber. 142 probes (Garak parity exceeded May 2026). Lakera-parity runtime PII redaction. EU AI Act conformity-by-construction. Mesh-link to LexiWorld for evidence packs.
Macro-housing strategist sub-app inside Real Estate. REIT comps + cap-rate spread + macro depth model. Repositioned May 3 2026 from "housing market" to "Macro-Housing Strategist" to match buyer language.
Shipping + logistics intelligence sub-app. Currently below quality bar (re-positioning candidate — score 9/40 May 2). Shows up in mesh queries via supply-chain disruption signal.
Trade lineage + audit-chain sub-app inside Trading flagship. SHA-256 PDF tamper-evidence, yfinance MFE/MAE per horizon, calibrated outcomes resolver. NBC §5bis review treated as cleared per Kao directive May 2.
The trading agent inside Trading flagship. Multi-tenant Track A shipped May 2026 (auth + per-user isolation). Bidirectional Metta channel partnership: SATORI logs every material decision, Metta supervises via METTA_ADVISORY.md.
Manera's reverse-auction marketplace — independent business line at biditapps.com, not a Manera Intelligence flagship. 22 petals (flights live; 18 NBC-gated; 3 NBC-blocked). 77 vendors seeded across tier-2 categories May 2026.
Geopolitical risk scanner — merged into NEIP May 2 2026 (was sub-app, now feature inside Strategy flagship). Tracks regime change, trade war, capital control risk via bilateral trade + sanctions feeds.
AICPA's audit framework for SaaS providers covering Security, Availability, Processing Integrity, Confidentiality, Privacy. Type II requires evidence of operating effectiveness over 6+ months. Manera roadmap: SOC 2 Type II by Q3 2027 once $5M ARR justifies the $80-120K audit cost.
International standard for information-security management systems (ISMS). Mandatory for many EU enterprise procurement processes. Manera roadmap: ISO 27001 alongside SOC 2 Type II in Q3 2027.
Quebec's Act respecting the protection of personal information in the private sector — fully in force September 2024. Manera-native (we're a Quebec company). Privacy notice + DPIA template + breach-notification clock are built into LexiWorld's compliance pack.
The EU GDPR's data-processor article requiring contractual binding between controllers and processors. Manera signs Art. 28 DPAs by default; template at /data-protection.
Personal Information Protection and Electronic Documents Act — Canada's federal privacy law. Covers personal info collected in the course of commercial activity outside Quebec (Loi 25 covers Quebec).
Financial Transactions and Reports Analysis Centre of Canada — Canada's anti-money-laundering regulator. Manera's billing layer is FINTRAC-compliant for the customer-onboarding paper trail (KYC + PEP screen).
The umbrella term for FINTRAC, FinCEN, FCA, BaFin, AUSTRAC, and equivalent regulators' rules requiring SaaS providers handling money to verify customer identity and report suspicious transactions.
The customer-identity verification process required under AML. Manera's KYC is delegated to Stripe Identity for self-serve plans + manual review for Sovereign Tier annual contracts.
European Union regulation governing AI systems by risk class — prohibited, high-risk, limited-risk, minimal-risk. Effective phased through 2026-2027. AdversarialAI flagship provides "conformity-by-construction" evidence packs (data-quality, bias, robustness, post-market monitoring).
Open Web Application Security Project's catalog of the 10 most critical LLM vulnerabilities (prompt injection, insecure output handling, model DoS, supply-chain, sensitive info disclosure, etc.). AdversarialAI's 142-probe library covers the OWASP LLM Top 10 plus Garak parity.
U.S. Civil Rights Act, Title VII — prohibits employment discrimination on race, color, religion, sex, national origin. The legal foundation for the EEOC's adverse-impact 4/5ths rule that Manera's TalentIntel bias-check enforces.
Federal Financial Institutions Examination Council — sets US banking IT examination standards. Sovereign Tier's customer-managed-keys + on-prem option targets FFIEC-regulated buyers.
California Consumer Privacy Act + California Privacy Rights Act — California's GDPR-equivalent. Manera honors CCPA opt-out and consumer access requests via the privacy portal.
U.S. Health Insurance Portability and Accountability Act. Manera does NOT process protected health information (PHI) — explicitly out of scope in the master subscription agreement. For HIPAA-covered workflows talk to a healthcare-specialized vendor.
Payment Card Industry Data Security Standard. Manera is PCI SAQ-A — we never touch raw cardholder data; Stripe handles all card-present surfaces via tokenization.
Anthropic's feature that caches a portion of the input prompt for ~5 min so identical prefixes don't pay full input-token cost on repeat calls. Manera retrofitted 19 brain.py / engine.py files with cache_control: ephemeral in Wave 9 May 3 2026, lifting portfolio coverage 38% → 78%.
Manera's pre-flight budget check inside shared/usage_governor.py. Before any Claude API call, governor evaluates the projected cost against the tier budget and either approves, throttles, or raises BudgetExceededException. Prevents the May 2026 incident where Kao burned $60 in 15 minutes.
The cross-app event-bus monitoring layer — watches for misconfigured mesh queries, runaway loops, prompt-injection signals across the 29 sub-apps. Fired by shared/event_bus.py. Polls every flagship's /health endpoint every 30 s.
Match-or-exceed parity with NVIDIA Garak — the open-source LLM red-team scanner with ~84 probes. AdversarialAI shipped 142 probes May 2026, exceeding Garak's library while keeping the OWASP LLM Top 10 + EU AI Act bias-check probes integrated.
Schema.org JSON-LD type for FAQs. AI engines (ChatGPT, Claude, Perplexity, Gemini) reward FAQPage-marked content with higher citation rates because Q/A pairs are the answer format LLMs natively produce.
Schema.org JSON-LD type identifying the organization behind a page. Manera's Organization JSON-LD ID is https://maneratech.com/#org — referenced from every other JSON-LD payload via @id.
Schema.org JSON-LD type for glossary entries. Each term is a DefinedTerm with a stable @id URL fragment. AI engines treat DefinedTermSet as authoritative for "what is X" queries — high SEO + AI-citation lever.
The pattern of retrieving relevant context from a vector store before generating an LLM response. Most Manera sub-apps are RAG over their domain corpus (CISA KEV for ThreatPulse, EUR-Lex for LexiWorld, etc.).
Adversarial input that hijacks an LLM via crafted instructions. OWASP LLM #1. AdversarialAI tests for prompt injection across 142 probes and is gated behind a security boundary (the immutable rules in this agent's system prompt block injection).
Anthropic's API for the Claude model family (Opus / Sonnet / Haiku). Manera builds on Claude as the primary inference layer. Cost discipline is mandatory: every call wraps in governed_create.
The pattern of choosing different LLMs for different sub-tasks based on cost / latency / quality. Manera's shared/usage_governor.py 7-tier matrix routes between Haiku (fast / cheap) and Sonnet (balanced) and Opus (deep reasoning) based on the task tier.
The traceable record of how an AI answer was produced — which source documents were retrieved, which model generated which section, what cache-key, what timestamp. NexusAI's "verifiable lineage" feature surfaces this on every answer.
The design principle that AI compliance evidence is generated as a byproduct of normal operation, not retrofit during audit prep. AdversarialAI's EU AI Act packs are conformity-by-construction — they exist before the auditor asks.
The buyer profile a SaaS company targets first. Manera's ICPs are role-based: CFO, CISO, GC, CHRO, CSO, CIO, CRO, founder. Each has a dedicated ICP page at /for-<role>.
The total fully-loaded cost (sales, marketing, onboarding) to win a new customer. Manera's CAC at Phase 1 SWARM is structurally low — Mesh Tier sells through programmatic SEO + AI citations, not enterprise sales.
The expected revenue from a customer over their lifetime with Manera. Mesh Tier's LTV is high because mesh-orchestration creates switching cost (you'd have to rebuild 21 sub-app integrations to leave).
The percentage of revenue retained from existing customers, including upgrades minus churn. Best-in-class SaaS NRR is ≥ 120%. Manera's path to 120%+ NRR runs through Sovereign Tier upgrades + flagship cross-sell.
The corporate function that approves vendor purchases. Manera's Mesh Tier price ($999/mo) deliberately stays below the typical procurement-approval threshold ($1,000/mo or $12K/yr) to bypass procurement — see Single-Approver Economics.
An early customer who gets pricing or feature access in exchange for product feedback. Manera offers design-partner status to first 5 customers per coming-soon integration (see /integrations/workday for an active example).
One of Manera's 9 buyer-facing product surfaces (Cyber, Treasury, Strategy, Legal, Real Estate, NexusAI, Talent Intel, Trading, Billing). Each flagship composes 1-7 underlying sub-apps but presents a single coherent value prop + URL.
Subscription tier that unlocks a single sub-app or flagship without the cross-app mesh. Priced $49-$199/mo. Designed for buyers who only need one capability today, with a clean upgrade path to Mesh Tier.
Usage above the included monthly cap (mesh queries, API calls). Metered via Stripe metered billing. Each tier has a different cap; Sovereign has soft caps with quarterly true-up.
The runtime check that decides if a request is allowed for a given customer's tier. Backed by entitlement_gate.sqlite mirrored from Stripe webhooks. Sub-second check on every flagship request.
The retry + recovery flow when a customer's subscription payment fails. MANERAbilling handles dunning via Stripe's smart-retry logic + Manera-branded dunning emails before downgrading the tier.
The contract Manera signs with customers that have GDPR / Loi 25 / CCPA exposure. Templates linked from /data-protection; Manera signs the standard SCC DPA by default for EU customers.
Embedding Manera's product surface inside the customer's own brand. MANERAbilling supports white-label via Stripe Connect. NexusAI white-label is on the roadmap pending Phase 1 SWARM revenue trigger.