Answer
Phrase CAT: What the Editor Does Inside a Translator's Workflow
Phrase CAT is a cloud-first segment editor with live TM, term-base hits, inline QA, and AI pre-translate — priced $27–$55 per user per month on professional
The Phrase CAT editor is the segment-level production surface inside a wider cloud language platform, where translators work through source text one segment at a time against live translation memory, term-base hits, inline QA, and AI pre-translate scores. This page sits inside the Computer-Assisted Translation Tools: The Working Translator's Field Guide, which maps the 7-tool professional CAT cluster Phrase belongs to.
Key points
- Phrase CAT is the browser-based segment editor inside the Phrase Language Intelligence Platform, formerly branded Memsource until the 2021 rebrand — it runs entirely in a browser with no installation required, though a downloadable desktop editor for Windows and macOS is also available.
- Phrase is priced at
- Phrase CAT is the browser-based segment editor inside the Phrase Language Intelligence Platform, formerly branded Memsource until the 2021 rebrand.
- Phrase is priced at $27–$55 per user per month on professional tiers, with the CAT editor bundled into Phrase TMS seats.
- Phrase surfaces TM match bands at 75–84%, 85–94%, 95–99%, 100% exact, and 101% in-context against a 75% fuzzy floor, each discounting the $0.10–$0.25 per-source-word base rate.
- Phrase AI pre-translate scores MT output per segment so high-confidence segments go to light post-edit and low-confidence ones go to full human translation.
- Phrase CAT output still has to pass an LQA gate priced $40–$90 per 1,000 words, scored against an MQM-style scorecard aligned to ISO 17100.
7–$55 per user per month on professional tiers, with the CAT editor bundled into Phrase TMS seats rather than sold as a standalone product; enterprise buyers negotiate custom pricing above those list tiers.
- Phrase surfaces TM match bands at 75–84%, 85–94%, 95–99%, 100% exact, and 101% in-context against a 75% fuzzy floor, and each band discounts the $0.10–$0.25 per-source-word base rate that governs certified translation production.
- Phrase AI pre-translate scores MT output per segment so high-confidence segments route to light post-edit and low-confidence segments route to full human translation, making the score a direct pricing lever on MTPE jobs.
- Phrase CAT output on regulated lanes still has to pass an LQA gate priced at $40–$90 per 1,000 words, scored against an MQM-style scorecard aligned to ISO 17100, which blocks publish until quality thresholds are met.
- The platform is built on infrastructure that maintains 99.9% uptime with zero-downtime deployments, and it complies with ISO 27001, GDPR, CCPA, and PCI DSS — making it viable for agencies handling sensitive or regulated content.
- Phrase TMS sits above the CAT editor layer and handles project routing, vendor assignment, and analytics, while multi-translator teams share a single TM in real time through role-based access controls.
What is Phrase CAT?
Phrase CAT is the browser-based segment editor inside the Phrase Language Intelligence Platform, formerly Memsource, where translators edit each source segment against translation memory, term-base hits, inline QA warnings, and Phrase AI pre-translate scores. A downloadable desktop editor exists for Windows and macOS, but the defaults are cloud-first. For the shared vocabulary — TM, term base, QA — see CAT Tool: What It Is, What It Does Inside a Translator's Workflow.
Buyers confuse Phrase with three adjacent things. The table below separates them in one glance.
| Term | What it is | When you need it |
|---|---|---|
| Phrase CAT editor | The segment-level bilingual grid where translation actually happens. | Any translator editing segments against TM and term base. |
| Phrase TMS | Project routing, vendor assignment, deadlines, analytics above the editor. | Agencies and localization managers running multi-vendor jobs. |
| Memsource | Legacy name for the same product before the 2021 rebrand. | Reading old contracts, CVs, or documentation. |
| Phrase AI | MT engine plus per-segment scoring layer feeding pre-translate. | Routing high-confidence segments to lighter post-edit. |
Is 'phrase cat' the editor or a cat idiom?
'Phrase cat' resolves to two unrelated queries: the Phrase CAT editor used by professional translators, and 'cat phrases' meaning idioms about cats. This page covers the editor only. The pet-vocabulary meaning is not addressed below.
Why is Phrase CAT sometimes called Memsource?
Memsource rebranded to Phrase in 2021 after acquiring Phrase GmbH, so legacy documentation, agency contracts, and translator CVs still refer to the same editor as Memsource. Vendor onboarding packs, PO templates, and job tickets issued before the rebrand use the Memsource name. The product and the file formats did not change.
What does the Phrase CAT editor actually do segment by segment?
The Phrase CAT editor splits a source document into segments and shows, for each one, the translation-memory match band, term-base hit, Phrase AI pre-translate score, and inline QA warnings in a single bilingual grid. The translator edits in that grid and the system writes the confirmed result back to the shared TM in real time, so every confirmed segment immediately benefits any other translator working on the same project or language pair.
A simplified pass through one segment looks like this.
Each segment routes through 5 decision bands: 101% in-context matches lock or confirm automatically; 95–99% matches go to light edit; 75–94% matches go to full post-edit; below-75% segments route to Phrase AI scoring, then either light MTPE (high confidence) or full human translation (low confidence). Every path ends with an inline QA check before the segment is confirmed and written to TM.[Light edit]
B -- 75-94% --> E[Full post-edit]
B -- Below 75% --> F{Phrase AI score?}
F -- High confidence --> G[Light MTPE]
F -- Low confidence --> H[Human translate]
C --> I[Inline QA check]
D --> I
E --> I
G --> I
H --> I
I --> J[Confirm segment, write to TM]
How does Phrase handle translation memory match bands?
Phrase surfaces TM matches against the industry-standard 75% fuzzy floor, with bands at 75–84%, 85–94%, 95–99%, 100% exact, and 101% in-context, each discounting the $0.10–$0.25 per-source-word base rate. The band drives the per-word payout the translator sees on the job sheet. A 101% in-context match is the highest confidence tier — it signals that both the segment string and its surrounding context are identical to a previously confirmed translation, so the segment is pre-confirmed or locked and billed at 0–25% of the base rate. A no-match or below-75% segment carries the full base rate because it requires original translation effort with no TM leverage.
| Match band | What it means | Typical rate treatment |
|---|---|---|
| 101% in-context | Exact match plus matching surrounding segments. | Often 0–25% of base rate. |
| 100% exact | Exact string match, context not confirmed. | Often 25–40% of base rate. |
| 95–99% | Near-exact fuzzy; small edits expected. | Often 40–60% of base rate. |
| 85–94% | Fuzzy; moderate edit effort. | Often 60–75% of base rate. |
| 75–84% | Low fuzzy at the floor. | Often 75–100% of base rate. |
| Below 75% / no match | Treated as new translation. | Full base rate, $0.10–$0.25 per source word. |
For a buyer-side comparison of translation memory software, see Choosing Translation Memory Software: What to Compare Before Buying.
What does the Phrase AI pre-translate layer score?
Phrase AI pre-translate runs the segment through an MT engine, scores the output per segment for confidence, and routes high-confidence segments to lighter post-edit while flagging low-confidence ones for full human translation. The score is not a surface-level estimate — it reflects the engine's internal confidence in the specific target output for that segment, accounting for domain, register, and term-base alignment. That score becomes the pricing lever in an MTPE job: lighter effort on scored-high segments bills at reduced MTPE rates, while scored-low segments carry full post-edit or human-translation rates. Phrase TMS records the pre-translate scores at the project level, giving project managers an analytics layer to audit routing decisions and refine thresholds over time. For deeper detail, see What Phrase AI Does and Whether It Fits Your Workflow.
How do term bases and inline QA show up in the editor?
Term-base matches highlight in the source segment and auto-insert on target focus, enforcing approved terminology without requiring the translator to search a separate glossary. Inline QA checks run continuously as the translator types, surfacing issues at the moment they occur rather than deferring them to a batch QA step. This means error correction happens at zero marginal cost — fixing a tag error mid-segment is a keypress, whereas catching the same error at final review requires reopening, re-confirming, and re-exporting the file. The inline warnings cover a comprehensive set of failure modes:
- Missing or reordered inline tags.
- Number and date mismatches between source and target.
- Segment length violations against UI character limits.
- Term-base divergence where an approved term was not used.
- Double spaces, trailing punctuation, and placeholder damage.
- Untranslated segments and copied-source warnings.
Comments and in-editor notifications complete the collaboration surface: translators, reviewers, and project managers can thread discussions on any segment without leaving the editor, and role-based access enforced through Phrase TMS determines which users can confirm, lock, or reject segments. The result is a single interface that handles translation, review, QA, and team communication in one place.
How it works
How does Phrase CAT fit inside a professional translation workflow?
1
How does Phrase CAT differ from Phrase TMS?
Phrase TMS owns project routing, vendor assignment, deadlines, and analytics across jobs, while the Phrase CAT editor owns segment-level translation inside each job — same platform, different layers.
| Layer | Owns | Primary user |
|---|---|---|
| Phrase TMS | Projects, vendor assignment, deadlines, analytics, role-based access. | Project manager, localization lead. |
| Phrase CAT editor | Segments, TM, term base, inline QA, Phrase AI pre-translate, comments. | Translator, reviewer, LQA reviewer. |
2
Which workloads does Phrase CAT fit best?
Phrase CAT fits continuous software-string localization, SaaS UI, help-center and documentation work, and repository-connected content — anywhere source files update constantly and multiple translators need shared live TM. Typical fits include:
- SaaS UI strings pulled from GitHub, GitLab, or Bitbucket connectors.
- Help-center and knowledge-base articles syncing from Zendesk or Intercom.
- Marketing site pages pushed from a headless CMS.
- Mobile app resource files on continuous release cycles.
- Product documentation stored as Markdown in a repo.
3
Where does Phrase CAT struggle on certified-translation lanes?
Phrase CAT's browser-first editing weakens offline control for certified lanes, where USCIS filings run $0.10–$0.25 per source word or $25–$150 per 250-word page and the certificate of accuracy is mandatory on every page. A downloadable desktop editor exists, but the defaults are cloud, and the certified-lane workflow needs tighter page-image control, signed PDF output, and offline archive copies.
Pricing
How is Phrase CAT priced for translators and agencies?
Phrase is priced at
Phrase is priced at $27–$55 per user per month on professional tiers, billed per seat, with the CAT editor bundled into Phrase TMS plans rather than sold as a standalone editor. The tier picks what sits above the editor: connectors, analytics depth, API quotas, and SSO.
| Tier shape | Price range | Typical buyer |
|---|---|---|
| Team / professional lower | About $27 per user per month | Small agencies, in-house teams starting a cloud workflow. |
| Professional upper | Up to $55 per user per month | Agencies needing connectors, analytics, and shared vendor pools. |
| Enterprise | Custom, above list pricing | Buyers needing SSO, SLAs, and custom API quotas. |
7–$55 per user per month on professional tiers, billed per seat, with the CAT editor bundled into Phrase TMS plans rather than sold as a standalone editor. The tier determines what capabilities sit above the editor itself — connectors to content sources, analytics depth, API call quotas, SSO, and the breadth of the shared vendor pool. Freelance translators working inside a client's Phrase TMS instance are invited as linguist users at no extra seat cost to the translator; it is the agency or enterprise buyer who holds the billed seats.
| Tier shape | Price range | Typical buyer | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Team / professional lower | About
Phrase is priced at $27–$55 per user per month on professional tiers, billed per seat, with the CAT editor bundled into Phrase TMS plans rather than sold as a standalone editor. The tier picks what sits above the editor: connectors, analytics depth, API quotas, and SSO.
7 per user per month | Small agencies, in-house teams starting a cloud workflow. | ||||||||||||
| Professional upper | Up to $55 per user per month | Agencies needing connectors, analytics, and shared vendor pools. | ||||||||||||
| Enterprise | Custom, above list pricing | Buyers needing SSO, SLAs, and custom API quotas. |
The platform's compliance posture is part of what justifies the per-seat cost for regulated buyers: the infrastructure is built to meet ISO 27001, GDPR, CCPA, and PCI DSS requirements, and operates at 99.9% uptime with zero-downtime deployments — a meaningful guarantee for agencies running continuous-localization pipelines where downtime translates directly to missed SLAs. For certified translation production work, the per-seat platform cost sits alongside the $0.10–$0.25 per-source-word translation rate and the $40–$90 per 1,000-word LQA gate, so buyers should model total cost of production across all three layers rather than evaluating the CAT seat price in isolation.
Computer-Assisted Translation Tools: The Working Translator's Field GuideHow does Phrase CAT compare to other professional CAT tools?
Professional translators cluster around 7 CAT tools in 2025 — SDL Trados Studio, memoQ, Phrase, Smartcat, Smartling, OmegaT, and Wordfast — and Phrase sits in the cloud-first, SaaS-string-heavy lane alongside Smartcat and memoQ cloud. The cluster is mapped in full Computer-Assisted Translation Tools: The Working Translator's Field Guide.
Phrase vs Trados: which fits regulated or certified work?
Trados Studio stays stronger on desktop control and offline certified lanes, while Phrase wins on continuous cloud workflows, shared live TM across vendors, and tighter repository-connector coverage. Trados Studio defaults to a Windows desktop install with local project packages; Phrase defaults to a browser session with cloud-stored TM. Side-by-side detail lives in Trados as a CAT Tool: Features, Costs, and Honest Trade-offs.
Phrase vs OmegaT: cloud SaaS or free desktop?
OmegaT is a free open-source desktop CAT that lacks live cloud collaboration, while Phrase charges $27–$55 per user per month for cloud TM sharing, Phrase AI pre-translate, and connector-driven continuous localization. OmegaT fits solo translators comfortable with manual TM and glossary files; Phrase fits teams needing shared state. The deeper read is OmegaT for Professional Translators: What It Does and What It Lacks.
How does Phrase CAT output pass certified-lane quality gates?
Phrase CAT output still has to pass the LQA gate priced at $40–$90 per 1,000 words translated, scored against an MQM-style scorecard aligned to ISO 17100 covering accuracy, fluency, terminology, locale convention, and style — the editor does not replace the gate. The LQA pass runs after the segment-level translation closes.
- Translator confirms all segments in the Phrase CAT editor, resolving inline QA warnings.
- Reviewer runs a second pass against TM and term-base enforcement.
- LQA reviewer scores a sample against the MQM-style scorecard aligned to ISO 17100.
- Error categories recorded under accuracy, fluency, terminology, locale convention, style.
- Score crosses the pass threshold or the job returns to the translator for rework.
- Signed certificate of accuracy issued per page for certified-lane deliveries.
A typical certificate of accuracy page, which Phrase itself does not generate, carries this shape:
| Element | What it says | Why it is there |
|---|---|---|
| Statement of competence | "I, [Translator name], certify that I am competent to translate from [source] into [target]." | Establishes translator standing for the receiving authority. |
| Statement of accuracy | "The attached is a true and accurate translation of the original [Document type]." | USCIS and similar bodies require this exact assertion. |
| Date and signature | ". Signed, [Translator name]." | Ties the certification to a dated act. |
| Contact block | "[Translator name], [Address],,." | Lets the receiving authority verify. |
Who should and should not pick the Phrase CAT editor?
Phrase CAT fits translators and agencies running cloud SaaS-string, help-center, and continuous-localization work with shared TM across distributed vendor pools; it fits less well for solo translators on offline certified lanes who need desktop-only control or who cannot sustain a per-seat subscription. The decision is not purely about features — it is about whether the surrounding workflow architecture matches what Phrase TMS provides above the editor layer.
- Pick Phrase CAT if the content stream is continuous, vendors are distributed, and connectors to GitHub, Zendesk, or a headless CMS matter.
- Pick Phrase CAT if Phrase AI pre-translate scoring will drive an MTPE pricing model.
- Skip Phrase CAT if work is predominantly offline certified filings priced $25–$150 per 250-word page with signed-PDF handoffs.
- Skip Phrase CAT if the budget cannot sustain $27–$55 per user per month per seat.
For an alternatives shortlist shaped around these verdicts, see Phrase Alternatives Worth Considering for Your CAT Workflow.
5–
How we work with Phrase CAT inside our certified-translation pipeline
We treat Phrase CAT as one segment surface among several, not as the gate. Our pipeline produces first drafts of certified translations, routes them to a human translator for post-edit against TM and term base, and then runs an LQA pass against an MQM-style scorecard aligned to ISO 17100 before issuing the certificate of accuracy. The editor choice follows the lane, not the other way round.
On cloud SaaS-string and continuous-localization work, Phrase CAT earns its seat on live TM sharing and connector coverage. On certified filings bound for USCIS or a court, we keep the signed-PDF and page-image control inside a desktop surface and use Phrase for the segment pass only.
We will not route a certified-lane job through a browser-only workflow and call the Phrase AI pre-translate score the quality gate. The score is a routing signal for post-edit effort, not a substitute for the LQA pass priced at $40–$90 per 1,000 words. We will not issue a certificate of accuracy unless a human reviewer has confirmed every segment against the source.
We also decline to re-price certified-lane work at MTPE rates on the basis of high Phrase AI scores. The $0.10–$0.25 per source word or $25–$150 per 250-word page floor exists because the receiving authority holds the translator accountable, not the engine.
Phrase CAT is the browser-based segment editor inside the Phrase Language Intelligence Platform, formerly Memsource, bundled with Phrase TMS on seats priced $27–$55 per user per month.
Frequently asked questions
What is the cat phrase?
In translation technology, "Phrase CAT" refers to Phrase's computer-assisted translation (CAT) tool — a professional editor environment where translators work through segments, apply translation memories, and manage terminology. "CAT" stands for Computer-Assisted Translation, a category of software that accelerates human translation rather than replacing it. Phrase is one of the leading platforms in this space, offering a segmented editing interface, glossary integration, and quality-assurance checks. For certified translation workflows, the CAT editor is where linguists review, edit, and finalize source-matched segments before certification.
Related
- Product Localization: How Translators and Agencies Deliver It
- Pseudo Localization: What Translators Need to Know About i18n Test Builds
- React i18n: How Translators Fit Into a React Localization Pipeline
- Service Localization: What Translators and Agencies Actually Deliver
- Software Translation Management: How Translators and Agencies Run a TMS Stack
- Squarespace Localization: How Translators Deliver Multilingual Squarespace Sites