Head-to-head · updated 13 September 2026
DataShield vs Castor: the catalog documents the data, but who governs the agent reading it?
First, the name. Castor became CastorDoc. On 19 March 2025, Coalesce bought it. They renamed it Coalesce Catalog. Did you search for Castor and land on a transform vendor's site? That's why. The product is good at what it always did well. Point it at a warehouse. It writes notes for every table and column. It maps column-level lineage. It lets a business user ask a plain English question, not file a ticket. Gartner named Coalesce a Visionary for metadata, in 2025. The hero line today is "Simplify data discovery for all data people." They earn it.
We're not a search tool for your whole estate. We won't say we are. DataShield Ontology is a governed data plane, with a catalog inside it. It scans and checks a live PostgreSQL source, in place. It tags columns against 129 PII, PHI, money and secret classes. It stores no column values, by default. We tokenize data sets at ingest. Agents query tokenized data over MCP. Turning a token back into real data is a locked step. It's checked every time. It runs on your own hardware. Here's the honest split, with the rows Castor wins too.
The short version
Pick DataShield when
- The catalog has to run inside your own network. Coalesce Catalog is SaaS. Private networking shows up only on the top tier. No self-hosted build is listed.
- You need the tool to find the PII nobody tagged. We tag columns against 129 classes. That covers all 18 HIPAA Safe Harbor IDs. We use regex, checksums and column-name clues. No model runs here, so the call stays the same each time.
- An examiner will ask you to prove an agent's access log wasn't edited. Our chain answers with math. Run the verifier.
- You want the data itself locked down, not just labeled. Tokenized at ingest. Broad-banded on the way out. Erasable, by wiping a key.
Pick Castor when
- You want every asset written up by Monday. Their AI notes are the best part of the product. We don't compete there.
- You need column-level lineage across Snowflake, BigQuery, Databricks, Fabric and your BI tools. Stitched end to end. Ours is typed and access-gated, but narrower.
- You already run Coalesce Transform. Catalog intent flows into the pipelines. Quality results flow back. That loop is real. We have nothing like it.
- The buyer is a business team, not a security team. Their search and helper are built for folks who don't write SQL.
Bottom line: Castor tells your people what the data means. DataShield decides what an agent may do with it. It keeps proof of that call, too. Most teams who buy us keep a catalog like theirs, for the rest of the estate.
Feature by feature: AI data catalog against governed data plane
Competitor cells state what Coalesce's public site, docs and changelog say. We use the date above. Did we get something wrong? Email support@myorg.ai. We'll fix it, and credit you.
| What matters | DataShield | Castor | Edge |
|---|---|---|---|
| Catalog coverage and connectors | We add a source over a live link. Then we scan and check it in place. No rows leave the source, ever. That's PostgreSQL today. Snowflake, BigQuery, Databricks, S3, Salesforce and Kafka sit on the roadmap. No code yet. We say so in the config file. | Snowflake, BigQuery, Databricks, Microsoft Fabric. Redshift, in private preview. Plus PostgreSQL, SQL Server, Tableau, Power BI, Looker, dbt and Fivetran. Much wider than us. | ◇ |
| AI documentation | We don't auto-write notes. We build our glossary from entity types, as ISO/IEC 11179 data element ideas. Precise, and dry. | The flagship feature. AI-built table, column and dashboard notes. Admin-set company context has fed the tool since October 2025. | ◇ |
| Lineage | Typed lineage, over a fixed word set. You can walk it, with per-hop access gating. A user only walks edges they may see on both ends. It comes from the pipelines that own the data, not a stored column graph. | Column-level and asset-level lineage. Snowflake Dynamic Tables support, too. Cross-warehouse lineage stitching, added in May 2026. Deeper than ours. | ◇ |
| Sensitive data classification | 129 built-in field classes cover PII, PHI, money data and secrets. Regex plus checksum checks: Luhn, NPI, Verhoeff, ABA, IBAN, GTIN. Plus column-name lists, and anti-pattern filters. Per-class risk level, confidence score and rule tie. Fixed, too. Two runs give the same call. | Manual tagging. Bulk column tagging landed in July 2025, to make it faster. We found no auto PII or PHI scanner in their public docs. | ◆ |
| Sample values in the catalog | Default rule is skip. We copy no column value into the catalog, at all. Masked stores shape only. Plain needs a clear click, and logs an audit row. A team lead can see just how many columns hold plain text. | Not listed in their public docs. Worth asking, since a catalog that stores previews of a PHI column has quietly become a second copy of it. | ◆ |
| Tokenization and masking | We tokenize data sets at ingest. Agents query tokenized data over MCP. Turning a token back is a locked step. It's logged, always. On top of that, broad bands for quasi-IDs. Dates become year, decade or age band. ZIPs become 3 or 4 digits. Phones, SSNs and emails get part masked too. Each carries a risk score. These are features you switch on, not defaults. | Tagging and access rules. No masking, tokens or blurring of values found. | ◆ |
| MCP and agents | More than 200 MCP tools, across Ontology, Auth, Corpus and Lighthouse. Agents don't just read metadata. They query the checked data, and act on it, under tool tokens with scope ceilings. | A Catalog MCP server since December 2025, with 8 read tools. Lineage tools added, May 2026. OAuth with per-user RBAC, since September 2026. Static API tokens for MCP are now dropped. Good hygiene. Read-only scope, though. | ◆ |
| Agent authorization | Every checked tool call passes a scope ceiling. It also passes an access tier, and a live check, before dispatch. Revoke an agent, and the next call fails closed. Not the next token swap. | OAuth, plus the platform's role-based access rules, for the signed-in user. Copilot has RBAC and audit logs. We found no per-call access check. | ◆ |
| Audit evidence | SHA-256 hash chain with Ed25519-signed checkpoints that are chained too. A check names the fault: tampering, insertion, deletion or a cut. Try the verifier. | Audit trails for the data lifecycle. Plus audit logs in Copilot. We found no public tamper-proof step. | ◆ |
| Break-glass and erasure | Scoped, time-boxed access for agents, just for a crisis, that auto-revokes on its own. GDPR erasure by crypto-shred of per-person key parts. Plus ISO 27560 consent receipts too. The chain still checks out after. | Not listed in their public docs. | ◆ |
| Deployment | Self-hosted, in your own cloud or room, your call. Or on a single-tenant box we run for you. Docker images cover Auth, Ontology, Corpus and Lighthouse. A signed deploy file, checked by Guardian. Ed25519 audit-signing keys sit on your box, in a key folder you hold. Kid-keyed turnover fails closed on an unknown key. KMS or HSM custody is a built gap in Auth, not a shipped link. HMAC tokenization keys sit as files on your box today, not in a KMS. | Multi-tenant SaaS. Private Networks show up on the Business Critical tier, along with a BAA. No self-hosted or on-site build appears in their docs. | ◆ |
| Maturity and pricing | Auth, Guardian and Lighthouse are all live now. Guardian and Lighthouse have run since April 2026. SOC 2 isn't in place yet, and we say so plainly. Published price and a quick quote. | Backed by a funded parent. Gartner Visionary, in 2025. Two firms bought and folded in. Developer tier is free. Starter runs $150 per user per month, billed each year. Catalog is gated by link count. It's not sold on its own. | ◇ |
◆ DataShield leads◇ Castor leads◈ comparable
We draw Castor claims from coalesce.io, castordoc.com and the Coalesce Catalog changelog. Last checked 13 September 2026. We link them below, not just from memory.
Three things you get here that you won't get from an AI data catalog
Labels the machine works out for itself
A tag is only as good as the person who thought to add it. We check values and column names against 129 classes, including the ones with a check digit. We filter the lookalikes, too. The result stays the same, from a config digest. No model sits in the path to drift. How the catalog works.
Authority that can change mid-flight
An analyst leaves on a Friday. Their agent is 20 minutes into a 40-minute job, unaware. DataShield checks the next checked tool call against live access, right then. It fails closed. A read-only helper would finish the job. It would tell you later. How Auth does it.
Proof that survives an audit
A log can be edited. That proves nothing. Ours is a hash chain with signed checkpoints instead. The verifier says what broke. Not just that it did. Two rules care about that most: EU AI Act Article 12 and HIPAA §164.312(b). Try it in your browser, no signup.
Where Castor is genuinely stronger
We'd rather you heard this from us. The CastorDoc team solved a dull, hard problem: getting a catalog written up at all. Their AI writes notes across the estate, and their search answers a plain question. Their lineage now stitches across warehouses, and into BI tools. Since the Coalesce deal, the story got tighter, not weaker. You set business intent in the catalog. The transform pipelines enforce it. Quality results from Coalesce Quality land back in the catalog. Is your estate Snowflake-shaped, with nobody sure what dim_cust_v3 means? Buy theirs. Their MCP work is sound, too. OAuth with per-user RBAC, shipped September 2026, beats most catalog peers. Dropping static tokens was the right call.
Here's the push-back, in short. Their MCP server hands a helper metadata, under the signed-in user's rights. That's a good read design. It's the whole design, too. There's no per-call access check. No break-glass. No tamper-proof record of what the helper did. Nothing touches the values themselves, either. Catalog governance is tagging plus access rules. That assumes someone already knew the column held a national ID. And all of it lives in their tenancy. Say your rule is that metadata about patient records can't leave your building. The talk ends before the demo starts. That's the seam we sell into. It sits underneath a catalog like theirs, quite happily.
Questions worth asking both of us
These are the questions we'd want answered if we were buying. Ask them on each call, ours included.
Can you cryptographically prove an audit log entry wasn't deleted?
DataShield: yes, it can. Each record ties to the one right before it. We sign and chain each checkpoint. A check tells delete from a cut from a tamper, each time. Try it at /verify. Castor: their site lists audit trails for the data lifecycle. Copilot logs its own moves, too. We found no tamper-proof step in their public docs. Ask them to show one.
What happens to a revoked agent mid-session?
DataShield re-checks access on each checked tool call, always. A pulled grant bites on the next call, no delay. Castor's MCP server signs the user in over OAuth. It applies the platform's role-based access rules, per request. That's better than a fixed token. Ask how fast a rights change reaches a helper mid-task. Ask, too, what record you get of the calls it made before the change.
How does GDPR erasure interact with the audit trail?
DataShield crypto-shreds the person's key parts. Then it sends an ISO 27560 consent receipt. Actor names in the chain carry HMAC locks, not real names. The proof still checks out, long after the person leaves. Castor lists no erasure step we could find. Worth asking. A catalog that stores sample values, or AI-written notes on a person's record, now holds copies of it, too.
I searched for a Castor alternative. Is that even the right search now?
Probably not. Here's the honest answer. Castor became CastorDoc. Coalesce bought it, in March 2025. It's now Coalesce Catalog, sold with Coalesce Transform and Coalesce Quality. The old castordoc.com still loads, with a banner saying the name changed. So the real question is this: do you want a catalog bundled with a transform tool? If yes, theirs is a strong bundle. If you wanted the standalone catalog you first shortlisted, look at us, Secoda or Select Star. Compare on what you actually need.
Can Coalesce Catalog run in our own data center?
Not that we can find. Their price page lists Private Networks and a BAA on the Business Critical tier only. No self-hosted or on-site build appears in their docs. DataShield ships Docker images for Auth, Ontology, Corpus and Lighthouse. A signed deploy file, checked by Guardian. It runs self-hosted, in your own cloud or room. Or as a single-tenant box we run for you. If self-hosting is a hard rule for you, ask them in writing, early.
Does DataShield have SOC 2?
Not yet. We won't say otherwise, ever. Auth is live now. It ships with a public threat model and a verifier anyone can run. Guardian and Lighthouse have both been live since April 2026. Design-partner terms include source escrow. That way one small vendor isn't a single point of failure. Details on the security page.
- Coalesce Catalog's current hero: "Simplify data discovery for all data people. Discover, centralize, and govern your data assets with help from AI." That's the pitch. - coalesce.io, 13 Sep 2026
- Coalesce acquired CastorDoc, and relaunched it as Coalesce Catalog: "an intuitive, AI-driven metadata management solution for modern data teams." Their own words. - Coalesce company news, 19 Mar 2025
- Catalog MCP server launched with 8 read tools, 8 Dec 2025. Lineage tools added, 15 May 2026. OAuth with per-user RBAC added, static API token auth for MCP deprecated, 7 Sep 2026. - Coalesce Catalog changelog, 13 Sep 2026
- Pricing: Developer free. Starter $150 per user per month, billed annually, with 3 Catalog integrations. Enterprise custom, with 5. Business Critical adds Private Networks and a BAA. No self-hosted edition listed. - coalesce.io/pricing, 13 Sep 2026
- Coalesce acquired SYNQ, and launched Coalesce Quality. That completed the Transform, Catalog and Quality "data operating layer." - Coalesce company news, 9 Mar 2026
- Coalesce was named a Visionary, in the 2025 Gartner Magic Quadrant for Metadata Management Solutions. - Coalesce company news, 24 Nov 2025
Other head-to-heads
DataShield vs Secoda
An AI workspace for metadata, and the enforcement it leaves out.
CatalogDataShield vs Select Star
Automated lineage, versus authority at the tool call.
CatalogDataShield vs Atlan
A metadata control plane, and the data plane underneath it.
AllEvery comparison
One honest scorecard per vendor.
See both moves run, right in your browser. Break a live audit chain. Revoke an agent mid-session. Then decide what your catalog still owes you. Demo Center access is free with a work email.
Get free Demo Center accessYou've seen the proof
Ready for a number? Scope your deployment and we'll price it against your own economics.
Get your quote →