FieldHash is the enterprise integration path.
Lumenais is the research hub. FieldHash.ai is the buyer-facing product surface for governed memory authority, answer-path evidence, and deployment conversations.
Answer path
The model only sees the governed packet.
Your retrieval layer can still retrieve broadly. FieldHash controls which retrieved records are allowed to become live influence, and which records remain visible only as blocked or caveated evidence.
1. App or agent
Your workflow asks a model to answer, act, summarize, decide, or continue a long-running task.
2. Candidate context
Existing memory, vector search, document retrieval, policy records, and prior decisions surface candidates.
3. FieldHash gate
Approved state is allowed forward; stale, rejected, superseded, or out-of-scope state is blocked or caveated.
4. Model of choice
The governed context packet is sent to the selected model endpoint: hosted, private, VPC, or local where configured.
5. Evidence packet
The answer path records what shaped the output, what was blocked, and which audit artifacts bind the event.
Integration modes
Start in shadow mode. Move to enforcement where risk is real.
The commercial path is deliberately practical: observe what your memory system would have sent, compare it with governed decisions, then enforce only the flows where stale or rejected influence creates review risk.
Typical pilot
One agent or workflow, one memory source, one model path, row-level traces, and an exportable evidence packet for technical review.
Pre-answer policy gate
Call Lumenais after retrieval and before generation. It returns the context packet, blocked records, governance reasons, and audit metadata.
Agent sidecar or SDK
Attach the control plane beside an existing agent runtime without replacing the whole application stack.
Audit sink
Export governed-memory events, checkpoints, blocked-record reports, and FieldHash-compatible evidence records for reviewer workflows.
Shadow-mode pilot
Start by observing candidate context and expected decisions without blocking production answers, then move selected flows into enforcement.
Deployment posture
Model-independent, not model-free.
FieldHash governs context before the model call. The generation model can be a hosted frontier API, a private endpoint, or a local/open-weight model route where configured. If an extraction or selection step uses a model, that step is explicitly recorded rather than hidden.
Hosted pilot
Fastest path for technical review, sample workflows, and benchmark-style trace inspection.
Customer VPC
Deploy the governance layer near customer data, with approved model routing and customer-owned logs where configured.
Private or on-prem
Run against private infrastructure and local model endpoints when data boundaries require it.
Existing memory stores
Use current retrieval systems as candidate sources. Lumenais governs influence; it does not require a rip-and-replace memory migration.
What Lumenais returns
A packet your reviewers can inspect.
Approved context passed to the model
Stale, rejected, superseded, or out-of-scope records blocked from influence
Governance reasons and scope metadata
Rollback and supersession links where available
Audit events, checkpoints, and FieldHash-compatible evidence where configured
Exports for review, SIEM, GRC, or internal evidence workflows where configured
Boundaries
What this does not claim.
Not legal compliance in a box.
Not a replacement for your retrieval database or document store.
Not training the base model weights during normal use.
Not a claim that every private memory, transient output, or internal trace is public or permanently signed.
Not a claim that Lumenais is smarter than the selected model at one-shot reasoning.
For enterprise integration, continue at FieldHash.ai.
Lumenais preserves the research lineage. FieldHash owns the commercial pilot, deployment, and buyer-facing integration path.
Request a FieldHash pilot