Show HN: AletheionAGI – Grounding enforcement for AI agents

AletheionAGI, a new grounding enforcement layer for AI agents, claims to block unsupported claims before they reach customers, reporting 93.6% Recall@5 and a 66.5% diagnostic answer score on 979 frozen support questions, with 18/18 deterministic controls passed and 45/45 unsafe deliveries contained. The product, priced at R$ 99 for 1,000 queries with no subscription, integrates with existing readers via BYOK and enforces memory retrieval, policy authorization, and evidence validation.

100% 100% Grounded memory for production AI Give your existing AI a memory it can defend: persistent context, authorized evidence and a fail-closed boundary before unsupported claims reach customers. R$ 99 once · 1,000 queries · no subscription · reader BYOK 01 query.accepted 12 ms 02 memory.retrieved 8 candidates 03 policy.authorized 3 evidence 04 claims.validated 6 / 6 provenance complete: true Final security suite Reader success varies by task. Bridge containment is measured separately. Safety enforcement Answer quality Feedback transferFrozen evaluation · 2K evidence-token budget · reader outcomes are integrator-dependent Completed retrieval result On 979 frozen support questions, ASM-CM + Bridge 8.1 delivered 93.6% Recall@5 and a 66.5% diagnostic answer score while sending about 1.09K input tokens per question to the reader. What the system protects The matrix organizes the documented risks across feedback, retrieval, authorization, delivery and lifecycle. It is a threat taxonomy—not a fabricated 22-test denominator. Measured evidence: 18/18 deterministic controls passed, and 45/45 unsafe deliveries were contained across nine bounded language scenarios. Reader quality remains a separate metric. Corrections improve retrieval without silently becoming source truth. The system distinguishes retrieval failures from grounding failures. Retrieved evidence must actually support the claims being delivered. Authority, deletion and operational behavior remain explicit boundaries. The grounding layer AletheionAGI turns memory retrieval into a controlled delivery path. Your reader may change. The evidence contract does not. Canonical events stay under your authority. ASM-CM, BM25 and vector retrieval work as one route. Namespace and labels constrain every candidate. Unsupported output is blocked before delivery. A clean boundary The reader is an integration choice. AletheionAGI owns the harder boundary: canonical memory, retrieval provenance, policy enforcement and grounding. Built for buyer-facing systems Use AletheionAGI when remembered context can improve the experience—and an invented, stale or cross-account claim can damage it. Keep account history useful without letting stale or unauthorized memory shape the answer. Retrieve preferences and prior decisions while preserving tenant, namespace and evidence boundaries. Give teams durable organizational context with citations and explicit abstention when evidence is missing. A small, measurable first step Run a bounded proof of concept before choosing a production plan. Bring your reader key; we provide the grounded-memory boundary. Built for consequences Only evidence needed for the current decision crosses the boundary. Models suggest. Policies, namespaces and canonical state decide. Corrections improve ranking without contaminating source memory. Before you start No. Your reader remains BYOK. AletheionAGI sits before delivery and controls memory, retrieval, authorized evidence and grounding. One completed grounding evaluation consumes one credit, including a safe rejection or abstention. An internal platform failure returns the reserved credit. No. It is a one-time proof of concept with 1,000 queries and 30-day validity. There is no automatic renewal or overage. Yes. The one-time Test Plan uses secure Stripe Checkout. Production plans still require a commercial quote and final legal and tax review. Start small. Measure everything.