Entify
github.com
- Free plan
- Web
- API
- Self-hosted
What Entify is
Entify is a free, open-source workspace for linking records, resolving entities and deduplicating datasets. It uses Splink 4 and DuckDB to match records and present clusters for review. Automatic setup identifies likely column types, suggests blocking rules and explains its decisions, which users can override. It reads CSV, TSV, Parquet, JSON and Excel files, and reports row counts, distinct values and completeness for each column. Matching options include exact, Jaro-Winkler, Jaccard and Levenshtein comparisons. Users can inspect match evidence and threshold analyses, export one surviving record per entity while retaining source IDs, and produce a PDF audit report. A Docker quickstart runs locally without accounts, API keys or external services, and uploaded data stays on the machine. Optional sign-in and shared storage use Clerk and Supabase; demo mode works without them. Entify can also run with Python 3.11 or later and Node 20 or later. The MIT-licensed project is aimed at data stewards and teams without specialist record-linkage expertise. Uploads have a 100 MB default cap, and the README describes the engine as memory-bound beyond roughly 100,000 rows. The backend API has no authentication and assumes localhost binding.
Who it suits
Entify suits data stewards and teams that need to link or deduplicate records without specialist record-linkage expertise. Local Docker operation may suit teams that want uploaded data to remain on the machine.
Good to know, and what to watch
- Suggests column types and blocking rules automatically.
- Reads CSV, TSV, Parquet, JSON and Excel.
- Provides match evidence and threshold analyses for review.
- Exports traceable merged data and PDF audit reports.
- Docker quickstart needs no accounts or external services.
- Uploads have a 100 MB default cap.
- The engine is memory-bound beyond roughly 100,000 rows.
- Backend API has no authentication.
- Only one resolution result is held in memory at a time.
Plans
Every plan and priceOpen-source
Open source
Getting started
- Open the Entify GitHub repository
- Run the Docker quickstart locally
- Use demo mode without sign-in, or connect Clerk and Supabase for optional sign-in and shared storage
- Upload CSV, TSV, Parquet, JSON, or Excel data
- Review suggested blocking rules and match evidence
- Export the merged data or a PDF audit report
Limits to check
Uploads are capped at 100 MB by default, and the README describes the engine as memory-bound beyond roughly 100,000 rows. Only one resolution result is held in memory at a time, so a second concurrent run replaces the first; semantic blocking also requires an extra installation with roughly 600 MB of Torch dependencies.
The Club's review
Overview
Entify is an open-source tool for linking records and deduplicating datasets. It is best suited to data stewards and teams that need guided matching without dedicated record-linkage expertise. Its local Docker option and traceable review workflow make it a sensible fit for smaller jobs, but its memory limits and unauthenticated API call for care.
Key features
Entify profiles columns by row count, distinct values and completeness, then suggests likely data types and blocking rules. It explains those recommendations and lets users override them, giving newcomers a starting point without hiding the matching choices. Splink 4 and DuckDB power exact, Jaro-Winkler, Jaccard and Levenshtein comparisons.
Reviewers can inspect match evidence, score and threshold analyses, and charts before accepting clusters. Merge exports keep one surviving record per entity and retain source IDs, which makes results traceable and supports undoing merges. PDF audit reports provide a separate record of the review. The tool reads CSV, TSV, Parquet, JSON and Excel files.
Semantic blocking is an optional extension that requires an additional installation and downloads roughly 600 MB of Torch dependencies. It is a substantial add-on for teams that need it, but unnecessary if the standard comparison methods are sufficient.
Entify can run locally through Docker without accounts, API keys or external services; uploaded data stays on the machine. Optional Clerk sign-in and Supabase storage support shared use. Local deployment is appealing for data control, but the backend API has no authentication and assumes it is bound to localhost. Do not expose it to untrusted callers as-is.
Pricing
Entify is open source under the MIT license. The Open-source plan costs nothing. It has a default 100 MB upload cap and holds only one resolution result in memory at a time, so a second concurrent run replaces the first. There is no paid tier to step up to in the plan described.
Platforms
Entify is available as a web app, API and self-hosted deployment. The Docker quickstart is the most direct local route; running from source requires Python 3.11+ and Node 20+. The 100 MB upload limit reflects full in-memory file loading, and the project describes workloads beyond roughly 100,000 rows as needing a DuckDB file backend and job queue. That makes Entify a poor fit for larger or concurrent production workloads without additional infrastructure.
Who it's for
Choose Entify if you are a data steward or a small team that wants explainable matching, guided setup and auditable merges on modest datasets. Its local Docker path suits users who prefer data to stay on their own machine. Look elsewhere if your workload exceeds the stated memory range, needs concurrent runs, or requires an API that can safely face untrusted callers without added security work.
Pros and cons
- Pros: Automatic configuration explains suggested blocking rules and allows overrides, helping non-specialists get started while retaining control.
- Pros: Evidence review, source-ID-preserving exports and PDF audits make merge decisions easier to examine and trace.
- Pros: Local Docker use needs no account, keys or external services, and the project is MIT-licensed.
- Cons: The 100 MB default upload cap and full-memory processing constrain dataset size; workloads past roughly 100,000 rows need a file backend and job queue.
- Cons: Only one result stays in memory, so a concurrent run replaces the prior one.
- Cons: The API lacks authentication and should not be exposed to untrusted callers as-is.
- Cons: Optional semantic blocking adds roughly 600 MB of Torch dependencies.
Alternatives
For a sandbox or demo portal below 10K records, Koalify offers a free forever plan and a web interface; choose it when that hosted, smaller-scope setup fits better than Entify's local workflow. Plauti Deduplicate has a free Salesforce version, making it the more relevant starting point for Salesforce users. Senzing Entity Resolution has a 10M-DSR subscription priced at $58,560 a year, a markedly different paid option for readers considering that stated scale and budget.
Tilores offers Studio Local Mode for up to 100,000 records per local deployment with no account or licence key, making it an alternative for local workloads near Entify's stated range. Workbooks is a web, iOS and Android option with a 30-day free trial and a CRM plan, rather than a focused record-resolution tool. AWS Entity Resolution offers provider matching priced per 1,000 records processed, with provider licensing or subscription costs additional; consider it when provider matching is the need. DQE One Unify offers SaaS curative processing hosted in DQE's cloud, a different fit for teams seeking cloud-hosted processing. MatchCore Standard Edition costs $417 a month, billed annually at $5,000 per user license, and supports up to 1M records per matching run; it is the alternative to consider for larger stated runs if per-user licensing suits.
Browse more options in Data Deduplication Software.
Verdict
Entify is a strong starting point for teams that want free, local, explainable record matching and exports they can trace. Its guided setup and review workflow are the main reasons to choose it. The 100 MB cap, limited concurrency and unauthenticated API are decisive reasons to look elsewhere for larger, shared or exposed deployments.
The verdict
Entify offers guided record matching, review and traceable exports, with a local Docker path. Its upload, memory and API-security limits matter when planning deployment or larger workloads.
Questions people ask
Does Entify cost anything?
No. The open-source plan is free at no cost.
Which file formats can it read?
Entify reads CSV, TSV, Parquet, JSON, and Excel files.
Can it run without accounts or external services?
Yes. The Docker quickstart needs no accounts, API keys, or external services, and demo mode works without optional sign-in.
Where can Entify run?
It is available as a web app, API, and self-hosted software, with local Docker or Python and Node setup.
Does the API require authentication?
The backend API has no authentication and assumes it is bound to localhost.
What matching methods are available?
It supports exact, Jaro-Winkler, Jaccard, and Levenshtein comparisons, with semantic blocking as an optional extra.
What the maker publishes
- API security
- The README says the backend API has no authentication and assumes it is bound to localhost. source
- Automatic setup
- Its auto-configuration identifies likely column types, proposes blocking rules, and explains decisions that users can override. source
- Automatic setup
- Its auto-configuration reads CSV columns, infers their roles, proposes blocking rules, and explains its decisions. source
- Concurrency limitation
- Only one resolution result is held in memory at a time, so a second concurrent run replaces the first. source
- Data formats
- It reads CSV, TSV, Parquet, JSON, and Excel files. source
- Deployment
- It can run locally with Docker or with Python 3.11+ and Node 20+; the README identifies FastAPI, Splink 4, DuckDB, and Next.js in its stack. source
- Export
- Its merge export produces one surviving record per entity and retains source IDs so merges can be traced or undone. source
- File formats
- It reads CSV, TSV, Parquet, JSON, and Excel files. source
- Integrations
- Optional sign-in and shared storage use Clerk authentication and Supabase persistence; demo mode works without them. source
- Integrations
- Optional sign-in and shared storage use Clerk authentication and Supabase persistence. source
- Intended users
- The project describes itself as intended to make record linkage usable by data stewards and teams without specialist record-linkage expertise. source
- License
- The repository lists an MIT license. source
- Limits
- Uploads have a 100 MB cap, and the README describes the engine as memory-bound beyond roughly 100,000 rows. source
- Local operation
- The README says the Docker quickstart needs no accounts, API keys, or external services, and uploaded data stays on the machine. source
- Local operation
- The README says the default Docker quickstart requires no accounts, API keys, or external services, and uploaded data stays on the machine. source
- Matching
- It supports exact, Jaro-Winkler, Jaccard, and Levenshtein comparisons, with semantic blocking available as an optional extra. source
- Matching engine
- It uses Splink 4 and DuckDB to match records and review explainable clusters. source
- Optional semantic blocking
- Semantic blocking is optional and requires an additional install that downloads roughly 600 MB of Torch dependencies. source
- Profiling
- It reports row counts, distinct values, and completeness for each column. source
- Purpose
- Entify is an open-source workspace for resolving entities, linking records, and deduplicating messy datasets. source
- Purpose
- Entify is an early open-source workspace for entity resolution, record linkage, and data deduplication. source
- Review and audit
- It offers score and threshold analysis, match evidence charts, and PDF audit reports. source
- Review and export
- Users can inspect match evidence and threshold analyses, export deduplicated data, and generate a PDF audit report. source
- Scale limitation
- Uploads are read fully into memory with a 100 MB default cap, and the README says workloads past roughly 100,000 rows need a DuckDB file backend and job queue. source
Checked against the maker's pages on 9 Oct 2026.