Company Brain for people and AI agents
ValorBrain is the shared memory that keeps work moving when the person, the agent, or the model changes. Decisions, rules, and handoffs get captured in the flow and handed to whoever comes next, with source, permission, version, and history.
Internal benchmarks, not peer-reviewed; every number states its metric, reader and run. See caveats
Documents and vectors in Brazil · LGPD · isolation between customers
Context unit
The problem
With three or more agents in operation, context fragments faster than any wiki can keep up. The cost never shows up as a budget line: it shows up as repeated briefings, outdated answers, and rework.
Fragmented context
“Where was that decided, again?”
Repeated briefings
“Let me paste the doc again.”
Stale documents posing as answers
“Didn't this rule change?”
Permissions rebuilt from scratch
“Is this agent allowed to see finance?”
Same team, two context architectures
ADR-007 buried in chat history
test convention pasted by hand
has no idea ADR-007 exists
briefing rebuilt from zero
March's INFRA.md
no access to team decisions
re-explains everything in standup
reviews permissions tool by tool
Four partial views of the same work. No current version anywhere.
How it works
No modeling project before you start. ValorBrain captures what already happens in agent sessions and team tools, keeps it with source and version, and hands it to whoever needs it next, with the permissions of whoever is asking.
Hooks in the agents' cycle record decisions, open items, and handoffs while the work happens. Documents, transcripts, and connectors complete the picture.
Every context unit has a type, a source, an owner, and a validity window. When something changes, the previous version stays in the history, and conflicts between sources surface before they become an answer.
At question time, context reaches the agent through MCP, the REST API, or a handoff, ranked by what matters to whoever is asking. The answer cites its source.
“Let's standardize cursor pagination for public APIs. Offset is superseded.”
Decision recorded as CTX-2026-ENG-007 · v1, with source in ADR-007.md and PR #412.
“Create the invoice listing endpoint.”
Endpoint created with cursor pagination, per ADR-007.
[Fonte: CTX-2026-ENG-007 // v1]
Nobody re-explained the pattern. Cursor started from the previous day's decision.
Governance
Per-person and per-source permissions applied at the semantic layer. The agent inherits access from whoever uses it; sensitive domains stay isolated; every read is traceable. When context is missing, the answer says so and points to the person responsible.
Simulator · same question, different access
Illustrative exampleQuestion asked to the agent
“What's the maximum discount without executive approval?”
10% without executive approval. Above that, RevOps validation is required.
[Fonte: CTX-2026-GTM-904 // v3]
▸inherit(user=marina)
→scope: gtm/pricing ✓
→CTX-2026-GTM-904 · v3 · current
→read-log recorded
Anything not ready yet is marked there as in preparation.
Proof
If the right context doesn't arrive, the agent starts over. High recall is the precondition for continuity; the final answer also depends on the model that reads the context.
LoCoMo
per-session recall@10 · random 10-session draw: 42.25%
96.58%
BEAM-100K
end-to-end answers · average (binary 80.8%)
75.5%
Benchmark configuration: reader GLM-5.3 Flash (Ox Alpha) at effort=max — a benchmark configuration, not the product default (the reader is the customer's LLM; ~44% slower than default effort); judge GLM-5.2; memory through the production path (memory_prepare); 20 conversations, 400 questions; N = 1 run; run 2026-08-23
Read this before citing
Verticals
The mechanism is the same; what's worth remembering changes. Each vertical comes with a scenario, typical activities and the indicators that make sense to track. No invented targets.
Featured department · Service & Support
Pricing
A fixed, published monthly fee. Nothing is billed per response and there is no document quota.
US$0
1 person · one project · no card
US$39/mo
1 person
US$99/mo
5 people included
US$199/mo
20 people included
From
US$999/mo
scope by contract
Questions
No “absolutely!” and no miracle promise. If your question isn't here, it's a good way to start a conversation.
A wiki is static. ValorBrain versions, cross-references sources, injects context into the agent, and surfaces a conflict when two sources disagree. The wiki can stay, as one source among others.
RAG answers a question. ValorBrain keeps decisions, permissions, versions, and handoffs between agents, and knows when an answer went stale because the rule changed.
No. It lives in the hooks and MCP of the agents the team already uses. Capture happens during work and delivery at question time, with no extra meeting and no new screen.
Documents, vectors and the database stay in Brazil. Answer generation uses a model provider outside the country, and the encrypted backup lives in the cloud (Google Drive). In compliance with the LGPD. Tenant isolation, permissions inherited per person and source, read auditing, and no outbound calls from the engine. The connectors live in the SaaS.
Every rule has a version, a validity window, and an owner. The current version arrives before the old one, conflicts between sources and outdated documents get flagged, and material changes go through human review before they circulate. What we promise is traceability: every answer shows which version it came from.
No. Migration happens without rewriting the past: existing documents, decisions, and transcripts come in through ingestion, and versioning starts from there.
Automatic hooks do the capture; memory types (decisions, handoffs, lessons) add structure; the memory health report and auditing show what has aged. Human curation goes where it makes a difference: validation and review.
Because every agent that joins without context means starting over again, every time. With one shared memory, each new agent starts from accumulated context, and the leverage compounds instead of diluting.
Next step
Bring a real project. In one conversation we map where context gets lost between the people and agents on your team, and where to start.