/
Documentation
Molthood analyses tokens, wallets, contracts, and websites on Robinhood Chain, and reports what it could not check as carefully as what it could. 49 pages, all of them describing the system as it actually is.
Overview
What Molthood is, why it works the way it does, and what it runs on.
- About MolthoodAn AI execution platform for Robinhood Chain: you describe a subject, agents gather evidence, and every finding carries its source.
- Why Molthood existsMost on-chain tooling answers confidently whether or not it checked anything. That gap is the product.
- ArchitectureHow a request becomes a report: routing, agents, evidence, and the summary written last.
- How live data worksWhere the numbers come from, when they are cached, and what happens when a source is unavailable.
- Supported chainsMolthood is built for Robinhood Chain. What that means, and what it does not.
- Supported APIsWhich capabilities the platform can call, and how it behaves when one of them is not configured.
Getting started
What Molthood does, how to get a key, and how to run your first analysis end to end.
- IntroductionMolthood is an execution platform, not a chat interface. It gathers evidence and tells you what it could not check.
- QuickstartCreate a key, run an analysis, and read the result — three requests, no SDK to install.
- AuthenticationHow keys work, what they are allowed to see, and the two separate limits that apply to them.
- ExamplesConcrete things to try, and what a good answer looks like.
- Best practicesHabits that make the difference between reading a report and using one.
Molthood Agent
The conversational surface over Molthood's analysis engine — what it can do and how to ask.
Core concepts
The five ideas the rest of the platform is built on. Read Evidence first — everything else follows from it.
- EvidenceFindings are confirmed, refuted, or unknown. A check that could not run never renders as one that came back clean.
- Execution modelOne request in, one report out: how routing, agents, and the pipeline turn text into findings.
- AgentsNine agents, six implemented. What each one collects, and what it needs to be able to run.
- Risk scoringHow a score is produced, why higher means safer, and what a ceiling score means.
- Providers and capabilitiesWhy the platform runs with zero credentials, and what adding a key actually changes.
- Research engineHow Molthood investigates something that is not on chain: a project, a site, a claim.
- Market enginePrice, liquidity, volume and holder distribution — and which of those actually tells you something.
- Wallet intelligenceWhat an address holds, how concentrated it is, and what cannot be seen.
- Token intelligenceContract powers, supply mechanics, and the questions that decide whether you can sell.
Guides
Task by task: analysing each kind of subject, watching one over time, streaming a run, and what the limits are.
- Analyse a tokenWhat a token analysis checks, what each finding means, and how to read a score that carries missing checks.
- Analyse a walletBalances plus a per-token risk screen, with an explicit ceiling when a check could not run.
- Analyse a contractVerification, source availability, and the owner powers a selector scan can prove without source.
- Audit a websiteDomain age, certificates, archive history, and page claims — the one analysis that needs no credential.
- Streaming a runRender the report about a third of the way in, and let the summary type in after it.
- Change detectionA second analysis of the same subject reports what moved — and which of those changes should alarm you.
- Watching a subjectRe-run an analysis on a schedule and get a change report each time. Off by default, and it spends your quota.
- Publishing a runExecutions are private. Publishing one is opt-in, per run, and redacts by selecting rather than removing.
- Limits and timingsQuotas, rate limits, caps, and how long each kind of analysis actually takes.
Platform
Security, privacy, limits, and the words used throughout these docs.
- Developer platformWhat the developer platform will be, why it is not open yet, and what to use instead today.
- SecurityHow credentials are handled, what runs where, and what Molthood deliberately cannot do.
- PrivacyWhat is stored, what is scoped to your key, and what never leaves your browser.
- Rate limitsTwo separate limits — one protects the server, the other protects your budget.
- GlossaryThe terms used throughout these docs, defined once.
API reference
Every route on the HTTP surface, with its authentication requirement and response contract.
- ConventionsBase URL, authentication, error shape, pagination — the things that are true of every route.
- AnalysisRun an analysis: by target, by free-form request, or as a stream.
- ExecutionsStored runs, permalinks, and the subjects derived from them.
- TasksOne request, one report: classify, resolve a workflow, run what is available, and say what was skipped.
- WatchesScheduled re-analysis of a subject, and the change report each check produces.
- KeysCreate a key, check your remaining allowance, and — with an admin key — manage others.
- Chain and agentsRead chain statistics and the live agent registry. No analysis, no quota.
- Providers and healthWhat is configured, what is missing, and whether the service can actually store anything.
- Public feedThe runs whose owners chose to publish them. The only unauthenticated view of execution data.
Roadmap
What is built, what is next, and what is only an intention.
FAQ
The questions people actually ask, answered without hedging — including the ones with uncomfortable answers.
- GeneralWhat Molthood is, what it is not, and what it will not tell you.
- Usage and billingQuotas, what counts against them, and what happens when you run out.
- Reading resultsUnknowns, missing summaries, empty change reports, and other things that look like bugs.
- Running it yourselfSelf-hosting, credentials, and what works with none of them.