Investigative Journalism Tools

Aleph, DocumentCloud, Datasette, Hunchly — the FOSS investigative stack used by OCCRP, ICIJ, ProPublica, and modern public-interest journalism.

The Panama Papers, Pandora Papers, and most modern document-leak investigations run on a small set of FOSS / non-profit tools. Pair with Civic Tech Overview, FOIA & Public Records, Court Records, OSINT & Reconnaissance, and Open Data (CKAN).

Document-investigation platforms (free / FOSS)

★ ★ Aleph — MIT; OCCRP's investigative-data platform; cross-references documents, leaks, sanctions lists, company registries, and structured data. What the Panama Papers / Pandora Papers / Russian Asset Tracker investigations run on. Self-hostable; OCCRP runs a hosted instance for member journalists. The default for serious cross-document investigation in 2026.
★ ★ DocumentCloud — free for verified journalists; non-profit (run by MuckRock News, Inc. since 2017); Document hosting + OCR + redaction + annotation + entity extraction; "embed a primary source in your story" tooling. The default for journalist document-publishing. Code largely FOSS on GitHub.
★ ★ Datasette — Apache 2.0; Simon Willison's project; publish a SQLite database as an instant interactive website with API. Used heavily for FOIA-released spreadsheets, campaign-finance data, government CSVs. Plugin ecosystem (datasette-cluster-map, datasette-graphql, datasette-publish-vercel). The fastest path from "I got a CSV via FOIA" to "here's an interactive public site."
★ OpenRefine (formerly Google Refine) — BSD; data-cleaning tool; clusters typo-variants, normalizes, reconciles to Wikidata. Indispensable for messy data.
★ Tabula — MIT; PDF-table extraction; JVM-based; the default tool for "rip tables out of FOIA-released PDFs."

★ Bellingcat's Online Investigations Toolkit — free curated list; the canonical OSINT-for-journalism reference.
★ GIJN (Global Investigative Journalism Network) Toolbox — free; investigative-journalism resource compendium.
See OSINT & Reconnaissance for the full OSINT stack (SpiderFoot, theHarvester, Maltego, Sherlock, etc.).

★ ★ Tesseract — Apache 2.0; the default FOSS OCR engine; reasonable accuracy on clean modern documents.
★ OCRmyPDF — MPL-2.0; CLI wrapper; turns scanned PDFs into searchable text.
★ PaddleOCR — Apache 2.0; modern OCR; handles handwriting / messy layouts better than Tesseract.
PDFPlumber, pypdf — Python; structured PDF extraction.

★ Followthemoney — MIT; OCCRP's data model for people-companies-relationships; the schema underneath Aleph.
★ OpenSanctions — non-profit; free under CC-BY-NC; aggregated sanctions / PEP / watchlist data; used by Aleph + many journalists. Commercial license available.
★ Wikidata — CC0; the canonical free entity database; reconciliation in OpenRefine targets Wikidata by default.
OpenCorporates — partial-free; corporate registries from many jurisdictions; rate-limited free tier; paid for bulk.
Sayari Graph — paid; commercial entity-resolution; used by the more-funded newsrooms.

★ ICIJ Offshore Leaks Database — free public extract from Panama Papers / Pandora Papers / Paradise Papers / Bahamas Leaks. Searchable.
★ DDoSecrets — non-profit; archive of leaked documents in the public interest.
WikiLeaks — historical archive; less active in 2024–26.

★ Datawrapper — free for journalists; charts, maps, tables; embedded in thousands of newsrooms. The single most-used data-viz tool in journalism in 2026. Paid plans for orgs.
★ Flourish — Canva-owned; charts + storytelling-shaped visualizations; freemium.
★ Observable Plot / D3 — BSD-3 / ISC; the build-it-yourself stack.
See Analytics for the broader analytics tooling.

Hunchly — paid; the dominant browser-extension OSINT-capture tool; hashes every page you visit during an investigation; chain-of-custody friendly. No FOSS equivalent of comparable quality.
Maltego (paid Pro/Enterprise; free Community) — see OSINT & Reconnaissance.
Skopenow / IRBSearch / IDI / Tracers / Pipl — paid; people-search; used in courthouse / locate work.
NodeXL — paid; social-network-graph analysis on Excel.
Palantir Gotham — paid (and politically loaded); intel-analyst platform; used by some investigative orgs but the politics is a story.

The FOSS investigative stack is genuinely complete: Aleph + DocumentCloud + Datasette + OpenRefine + Tabula + Tesseract + Datawrapper covers most public-interest investigations from raw documents to published story.
Hunchly is the biggest unfilled FOSS gap. Browser-extension chain-of-custody capture deserves a non-profit FOSS equivalent.
OCCRP's role can't be overstated. They build Aleph, run a hosted instance for journalists, maintain Followthemoney, and contribute to OpenSanctions. Donate.
AI + leaks is the 2024–26 frontier. Embeddings-based semantic search across leaked corpora is genuinely useful but needs careful epistemic framing. See AI / LLM for the underlying tooling.
Operational security matters. Journalists working with leaked documents should pair this stack with Anonymity & Privacy OS (Tails, Whonix) and SecureDrop for source intake.

You have a leaked document corpus to investigate: Aleph (self-hosted or OCCRP-hosted).
You need to publish a primary source alongside a story: DocumentCloud.
You FOIA'd a CSV and need to publish an interactive view: Datasette.
You have messy data that needs cleaning: OpenRefine.
You need to extract tables from a 500-page PDF: Tabula.
You want a chart in your story by 5pm: Datawrapper.
You're an investigative newsroom budgeting infrastructure: join OCCRP / GIJN / IRE; both share infrastructure costs.
You handle source materials at high risk: layer in Anonymity & Privacy OS.