1. Start from the job, not the brand name
The fastest way to choose the right page is to ask what job you are trying to complete. DeepBrainz has a public company site, product sites, model releases, and research pages. They are related, but they are not interchangeable.
If you are new, do not try to read every page first. Pick the product or page that matches the work in front of you, complete one useful action, and then follow the cross-links only when the next question is specific. That keeps the experience practical instead of turning product selection into research overhead.
- Need the agent-system map, company overview, or official positioning? Start at deepbrainz.com.
- Need autonomous knowledge work with persistent agent memory? Start at www.lexopedia.in.
- Need governed engineering agents, scoped execution, tests, and approval boundaries? Start at www.agentfoundry.in.
- Need public release names, release files, or support status? Start at huggingface.co/DeepBrainz.
- Need research context, evaluation direction, limitations, or labs-side release guidance? Start at labs.deepbrainz.com.
2. Use Lexopedia AI for autonomous knowledge work
Lexopedia AI is the current production surface for autonomous knowledge work in the DeepBrainz system. It is the right place when specialized agents should research, reason, synthesize, preserve memory, support planning, or prepare decision-ready structured outputs.
Use it before execution when the question is still ambiguous. A good Lexopedia session should leave you with cleaner background, stronger assumptions, and a better sense of what should happen next. If the next step is a code change, implementation plan, or repository task, move the refined intent into AgentFoundry.
- Good for turning a hard question into a clearer answer, memo, comparison, or next-step plan.
- Good for knowledge and engineering-adjacent work that benefits from agent memory, source context, and stronger structure than a casual chat interface.
- Not the right place if your main job is governed engineering-agent execution with approval and review checkpoints — that is AgentFoundry territory.
3. Use AgentFoundry for governed engineering agents
AgentFoundry is the governed engineering-agent layer. It is the right destination when specialized agents should plan, implement, verify, debug, coordinate, and prepare repository-aware handoff with visible constraints, tests, review notes, and human approval rules.
It is most useful when the expected output must survive review: a pull request, checked fix, codebase inspection, test result, rollback note, or a concrete explanation of why a proposed change should not be made yet.
- Use it when engineering agents need to produce software work that must be reviewed, not just researched.
- Use it when cost, reviewability, sandboxing, approval, and controlled autonomous delivery matter.
- Use Lexopedia first if the problem is still ambiguous and needs research before execution.
4. Use DeepBrainz-R and Labs when you need release and research guidance
DeepBrainz-R1 is the public research release family within DeepBrainz-R. Public release files and names live on the public release hub, while Labs explains research context, evaluation direction, release guidance, and limitations.
If a release question affects a real evaluation, verify the exact release file name and date instead of relying on a broad product description. Public release pages, model cards, and Labs guidance should be read together when support expectations matter.
- Use the public release hub for concrete public release files and model names.
- Use Labs for research guidance, evaluation direction, and what is exploratory versus supported.
- Do not assume every experiment or checkpoint is a production recommendation.
5. Use the Help Center when you are blocked or unsure
The DeepBrainz Help Center exists to reduce confusion across products and pages. Use it when you need product selection help, access support, troubleshooting guidance, model-page clarification, or a clean way to contact the team.