A discovery engine, not a chatbot.
Indago finds the chunk you're looking for — ranked, with a citation. No hallucination. No round-trip to a model in another country. No prompt becoming someone's training data.
- Deterministic. Same corpus, same query, same result — six months later. Audit-grade.
- Multi-corpus. Code, docs, papers, conversations — ranked together, or scoped with one keyword.
- Structural, not stringy. Found a 300× perf bug by matching shape, not text. Grep would have missed it.
- Local. Index, query, model — everything runs on your machine. Air-gapped works.
indago · session example// ranked, on-device, < 1 ms $ indago find "expensive computation per chunk" // 4 results, ranked by structural fit 1 · RollupNode.cs:537 complexity 217 2 · ExpressionOpType:2781 complexity 341 3 · ComputeNode.cs:1130 complexity 188 4 · StreamProvider.Events complexity 188 $ indago solve "why is GroupBy slow?" // evidence across code · papers · prior sessions → "110 GB/s practical ceiling for unsorted scatter-accumulate" ↳ MEMORY/groupby_perf_analysis.md ↳ paper · radix-aggregation · arXiv 2403.12891 ↳ session · 2026-03-21 · GroupBy perf sprint
A query, a ranked list, a citation. No prose, no guesswork.
Pick the field you work in.
A free core, then a bundle per audience — its own sources, scans and workflows. Thirteen verticals in all; six shown here, the rest listed below.
Map the patent landscape and freedom-to-operate.
USPTO · EPO · WIPO · Espacenet. Prior-art sweep, §102 / §103 ranking, claim charts you can file.
Build a systematic review that cites every trial.
PubMed · ClinicalTrials.gov · Cochrane. PRISMA + GRADE. No PHI leaves the laptop.
Audit a codebase with taint-flow analysis.
NVD · GHSA · KEV · OSV · EPSS. Secrets, injection, crypto, auth, IaC — every finding with file, line, and a fix. SARIF out.
Search ten years of your repo by meaning.
16 languages structurally indexed · dead-code · complexity · cross-domain bridges.
The clause you actually filed under.
EUR-Lex · CourtListener · SEC EDGAR. Clause & renewal scans, decision memos, validation binders.
Eight hundred papers, one search.
arXiv · OpenAlex · IEEE · Semantic Scholar. P-hacking, replication-gap and AI-paper scans.
+ seven more verticals — trademark, life sciences, civic, business, procurement, grants, design. One install; buy one, preview the rest. See all on indago.one ↗
Your data stays local. Your AI is your choice.
Everything indexes on your machine. What you connect on top is your call — six honest ways to keep your data yours, not one cloud-only deal.
Local data + public research.
Fetches public sources on demand. Your data, index and results never leave your machine.
+ a local LLM.
Ollama, llama.cpp or MLX on your machine writes the answer from ranked snippets. Locally.
+ your favourite cloud AI.
Claude, GPT, Cursor, Codex via MCP. Indago sends only the snippets it retrieves — never your data.
Full air-gapped.
No outbound network. Offline activation, no telemetry, no update check-ins. Nothing leaves.
Bring your own model.
Llama, Mistral, Qwen, a fine-tune of your own — Indago feeds it ranked snippets and never holds your keys.
Nothing phones home.
Telemetry is opt-in and aggregate-only; Sovereign mode disables it entirely. No usage data leaves the device.
Download. Index a folder. Ask a question.
Free Core ships the engine, the daemon, all nine discovery primitives and search across your files and conversations. Thirteen bundles activate from the same install — no SaaS sign-up, no telemetry, no seat minimum.