What it is

Daxtra is a CV parsing, semantic search and matching vendor that has sold since 2002 to staffing agencies, in-house TA teams and the software companies that build ATSs. It parses CVs and job descriptions in 40+ languages into 150+ structured fields (XML or JSON, over REST or SOAP, hosted or on-premise), and layers recruiter-facing products on top: Daxtra Capture and Magnet load and deduplicate CVs from inboxes and job boards into your ATS, Search Nexus runs semantic search and ranking across your database and online talent pools, Styler reformats and anonymises CVs in your branding, and Engage (from the April 2024 PivotCX acquisition) adds SMS, voice and video outreach. TalentFlow, launched 2025-03-18, bundles those into an agentic source-screen-engage flow, and Conversational Search (2025-05-21) replaced Boolean strings with plain-language queries. Strattam Capital has owned the company since June 2023, Michael Flannery became CEO in June 2025, and Daxtra reports 2,500+ staffing firms and hiring teams as customers.

The question this page answers: now that Bullhorn owns Textkernel, do you buy parsing and matching from an independent specialist, or take whatever your ATS ships natively?

Why it shows up in Recruiting stacks

It is the main independent parser left at enterprise grade. Bullhorn bought Textkernel on 2024-06-18, taking the other long-standing specialist, with 2,000+ customers including eight of the top ten staffing agencies, inside an ATS vendor. If you run a non-Bullhorn ATS, or you build recruiting software yourself, Daxtra is the established option whose roadmap no ATS vendor controls.
Language coverage is the product. 40+ languages with country-specific name, address and date handling matters when one desk places candidates from Poland, Japan and Brazil. An English-trained parser, or a generic LLM prompt, degrades on exactly the CVs a cross-border agency earns its margin on.
It plugs into the agency ATSs people already run. Daxtra has been a Bullhorn Marketplace partner since 2010 and is one of two search-and-match providers selectable inside Bullhorn Automation’s Auto Match. It also ships as a paid add-on in JobAdder, integrates with Vincere, PCRecruiter, Mercury, iCIMS and SAP SuccessFactors, and TalentFlow claims bi-directional sync with 40+ ATSs including Workday and Lever.

Pricing

Daxtra publishes no prices, offers no self-serve plan and no free trial. Everything is an annual contract on a quote. Pricing trackers describe the recruiter-facing products (Capture, Search Nexus, TalentFlow) as per-user, and the parser API is quoted on document volume.

Third-party estimates (ITQlick) put per-user products at roughly $75–$100 per user per month, falling with seat count, plus a one-off implementation fee of $2,000–$5,000 for a small agency. ooligo could not verify those figures against a real invoice, so use them as an opening anchor, not a budget.

For the parser API, the public market sets the ceiling. Affinda lists self-serve parsing at $0.10 per document and $3,600 a year for 66,000 documents (about $0.055 each). Textkernel’s Tx Platform charges 1 credit per parse, plus 0.1 credits for its LLM parser. If Daxtra’s parse-only quote lands well above $0.10 per document at your volume, you are paying for language accuracy. Make them show it on your own CVs.

Best for

Ops and TA-technology leads at multi-country staffing agencies (roughly 20–500 recruiters) who hire across three or more languages and want parsing, dedupe and semantic search on top of an ATS they keep. Also a strong fit for HR-tech product teams that need an embeddable, on-premise-capable parser that no ATS competitor owns.

Not for you if you hire in English only and run under about 5,000 CVs a month. A self-serve parser API, or a Claude extraction prompt with a JSON schema, covers that for $3,600 a year on Affinda’s list, or less. Also not for you if you’re already on Bullhorn and willing to pay for its Pro tier: Bullhorn’s own AI search and match starts there, and running a second vendor for the same job is duplicated spend.

Versus the alternatives

Textkernel (Bullhorn) — the other long-standing specialist, with 2,000+ customers, an LLM parser option and deep labour-market data. Pick it when you’re on Bullhorn or expect to be. The ownership question only bites if you compete with Bullhorn or run a rival ATS.
RChilli — also 40+ languages, sold mostly to the 1,600+ recruiting platforms that embed it, and running a dedicated Textkernel-migration page since the Bullhorn deal. Pick it when you’re an HR-tech vendor buying a white-label API, not an agency buying recruiter tools.
Affinda — the price-transparent entrant, with self-serve and self-hosted plans on a public list. Pick it when you need parsing only, want a number before a sales call, and can build search on top yourself.
DIY LLM extraction — the fastest-growing substitute. A structured-output prompt handles English CVs well at near-zero marginal cost. Pick it for a pilot or a single-language desk. It does not give you dedupe, skills taxonomies or ATS write-back.
Sourcing platforms — hireEZ and Eightfold sell search and match as part of a wider sourcing product. Pick one when the job is finding new candidates outside your database; Daxtra is stronger at mining the database you already paid to build. See the AI sourcing roundup.

Watch-outs

Integration quality varies by ATS. Capterra reviews (3.6/5 across 32) praise the dedupe and the Bullhorn fit but repeatedly report connector problems and slow fixes when it does not fit the customer’s system. Guard: make go-live acceptance contractual, with a named parse-accuracy and write-back test on 500 of your own CVs in your ATS, before the first annual invoice.
Accuracy claims are the vendor’s own. “Most accurate parser” and “95%+ accuracy” are Daxtra’s benchmarks, not independent ones. Guard: run a blind bake-off on the same 200–500 CVs, split by the languages you actually place, against Textkernel or RChilli and an LLM baseline, and score field-level accuracy yourself.
It is PE-owned and the product line is growing fast. A change of owner in 2023, the PivotCX acquisition in 2024, then a new platform and a new CEO in 2025 mean packaging will keep moving. Guard: sign 12 months with a written renewal cap, and list the specific modules and APIs you rely on in the order form so they can’t be re-bundled at renewal.
Parsed CVs are personal data at scale. Capture pulls CVs from inboxes and job boards into your database, which creates GDPR notice and retention obligations for every record, and AI ranking can pull you toward NYC Local Law 144 scope. Guard: confirm data residency and the DPA for the hosted service, set automated retention deletion in the ATS, and have counsel review Search Nexus ranking before you use it to decide who gets screened.

Edit this page on GitHub