Semvi by Xindar: Monitoring Scope, Plans and Buyer Verification

By Daoyu Guan · Product information reviewed October 31, 2026

Publisher disclosure: Xindar publishes this product explanation and supplies Semvi. The capability statements below distinguish Xindar's product-team confirmations, current public descriptions and checks still needed in a customer account. This is first-party product information.

Semvi is Xindar's AI visibility monitoring offering for teams investigating how AI answers describe and recommend a brand. Its intended business context includes B2B export marketing, manufacturers and agencies serving those companies. The practical buying question is whether a proposed subscription can produce the specific answers, source links and collection context that your team needs to inspect.

This page brings the product's identity, collection scope, plan structure and evaluation questions into one dated record. It also identifies the limits of the evidence behind those statements. A procurement team can use it to prepare a pilot and then replace outstanding questions with results from its own permitted account test.

Identity and official destinations

The current product name is Semvi. Use Semvi by Xindar where company attribution matters. Xindar remains the company and service brand. An existing URL containing geowatch can still lead to the current product explanation.

Use the legal contracting name stated in the order. This product explanation does not establish a registered trademark, an overseas office or an incorporation jurisdiction.

What collection coverage has been confirmed?

Xindar's product team confirms ChatGPT and Gemini consumer-web sampling, exports containing complete answers and citation URLs, and retention for at least one year. Longer retention follows the agreed service term; a three-year engagement can retain data for three years. The product explanation carries these attributed statements.

The public platform overview also describes API-based collection across ChatGPT, Claude, Gemini, Perplexity and Grok. Provider names and collection surfaces need separate fields in the proposal. An API run and a consumer-web run can represent different observation conditions even when the provider is the same.

Google Search, AI Overviews and AI Mode coverage remains unconfirmed. A Gemini collection statement does not establish support for those Google search experiences. Obtain the actual surface and mode for every requested combination before including it in a report or purchase order.

For each agreed collection stream, specify the following:

1. The exact question and its version.
2. The consumer product or API configuration being sampled.
3. The visible model or mode, where available.
4. The language, intended market and collection-location evidence that can actually be supplied.
5. The schedule, treatment of failures and retained answer-to-source relationship.

These are proposed procurement fields. Their inclusion here does not mean every field has been verified in a Semvi export. An unknown collection location should remain unknown; writing a country name into an English question does not establish that location.

How to read the plans

Consult the current public plan matrix for prices and billing periods. The matrix reviewed for this page lists the following allocation and frequency structure. Numbers refer to the published plan description rather than the results of a live allowance test.

Question allowances are shared across the plan's brands. A shared budget requires an allocation decision. For example, an agency assigning questions to three clients should record which client owns each question, how unused capacity is reassigned and what happens when a client leaves. Those operating rules should be agreed before interpreting a plan total as a delivery promise to each client.

A monitored question is also not automatically a billable run, an answer from every provider or an unlimited set of regional variants. Confirm how the selected collection surfaces and schedules consume the actual subscription allowance. The public totals alone do not resolve that accounting question.

This explanation publishes no additional price package. The public application is the price reference; a proposed order should identify its currency, billing period, included surfaces, taxes and any implementation work separately.

Agencies: brand capacity and client permissions need separate acceptance

Enterprise and Agency / OEM publicly describe team roles and multi-brand reporting. The public matrix also describes company-level data isolation. These statements are relevant starting points for an agency evaluation. They do not establish the permissions available between individual clients inside one agency company.

Use two permitted test clients and a restricted reviewer in the trial. Put a clearly labeled sample in each client's project. Check what the reviewer can open, search, share and export. Include direct links as well as the visible navigation. A permission result must concern the actual account configuration being purchased.

White-label reporting and resale terms require their own confirmation. A plan called Agency / OEM does not establish that all vendor branding can be removed, that every output can be redistributed, or that software access can be resold under an agency's own contract. Ask for an example of the proposed client-facing deliverable and the applicable rights.

This page does not report a passed client-isolation, white-label or resale test for Semvi. These are explicit evaluation items for buyers who need those workflows.

Follow one answer from the screen into a file

The product team confirms full-answer and citation-URL export. The public plans also describe CSV export and read-only access after a plan ends. A buyer still needs to establish the actual file contents, its link back to the screen and the duration of post-service access.

Choose one record that the account holder is permitted to inspect and export. Retain its exact question and observation time. Save the complete answer and the visible source URLs. Export that same record through the normal product workflow, then compare the screen and file.

Check five things:

1. Completeness: the file contains the full answer rather than a short summary.
2. Association: citations remain attached to the right question and answer.
3. Context: the file preserves the collection details promised in the order.
4. Missing evidence: failed collections and absent fields can be distinguished from valid answers with no brand mention.
5. Continuity: the agreed recipient can open the delivered file without depending on an undocumented account state.

These steps describe a proposed acceptance test. We have not completed that screen-to-file test in an authenticated Semvi customer account for this publication. Public demonstration pages and a statement that CSV export exists provide different evidence from an inspected customer file.

Write the retention start date, agreed duration, end-of-service access period and deletion rules into the service terms. At least one year of retention does not define every condition after a contract ends. Similarly, a public description of read-only access does not establish an unlimited post-service access period.

Interpret the dashboard through its units and time window

The public overview links to a demonstration dashboard. Its English interface, metric labels and help text were inspected on October 31. This establishes that selected public page's English presentation. Authenticated setup, reporting and export workflows were not tested for language completeness.

The inspection also found a wording issue worth resolving before using a headline figure: a card labeled “Total citations” had help text describing a lifetime count of answer runs containing a brand mention. Those are different measurement units. The visible accumulated count was not recalculated from underlying records, so this observation does not establish whether its stored value is correct.

A useful report names the unit and period beside the number. A mention-run count is the number of saved answers that contain a qualifying brand mention. A citation measure needs its own definition, such as displayed link occurrences, unique source URLs or answers citing a specified domain. An average recommendation position needs a rule for answers that mention a brand without placing it in an ordered list.

Ask which figures follow the selected date range and which are lifetime totals. For a percentage, request its numerator and denominator. Configured providers, providers with completed observations and providers where a brand appeared answer different questions. A missing observation should not silently become a negative answer.

These are interpretation and reconciliation requirements. This publication does not claim to have certified the dashboard calculations or tested their implementation.

Monitoring and implementation are separate scopes

Semvi supplies the monitoring offer. Xindar also describes English website and B2B export-marketing work. The proposal should identify the people and deliverables included in each workstream.

A monitoring finding may suggest that an English capability page needs a clearer process boundary. Turning that finding into a useful page still requires approved technical information, editing, engineering review and website implementation. A software allowance does not by itself purchase those activities.

For a manufacturer, assign an owner for technical claims and an owner for publication. For an agency, assign an owner for client approval and an owner for the monitoring panel. This keeps responsibility visible when a report identifies an inaccurate description or a source that needs correction.

What outcome evidence is available here?

This product explanation contains no verified customer revenue result, independent award or guaranteed AI placement. A sampled mention is an observation under recorded conditions. A useful recommendation additionally needs the correct entity and a relevant reason; visits, RFQs and sales require separate evidence.

For transparency, Xindar has published a separate five-day AI visibility research record. Its dated result is 73 ordinary valid consumer answers and zero qualified recommendations under the stated protocol. It includes an inspectable metadata projection and explains the limitations of the public material. It is first-party operational research, not a Semvi customer-account export or an independent validation of product performance.

A prospective buyer can therefore evaluate the stated product scope and request a permitted evidence demonstration. The material here does not justify assuming a particular improvement in recommendations or commercial results.

Questions before a pilot

Is Semvi a separate marketing agency?

Semvi is the monitoring product name. Xindar is the company and service brand. Use Semvi by Xindar when the relationship needs to be explicit.

Does each provider use the same collection interface?

The public API description and the product team's consumer-web statement describe separate surfaces. Specify the surface, mode and included frequency for the actual engagement.

Can an agency serve multiple clients?

Published higher-tier brand allowances provide a starting point. Test the required client permissions, evidence handoff and reporting rights before promising that workflow to clients.

Can we evaluate software separately from content work?

The application publishes monitoring plans and Xindar scopes implementation separately. Confirm the actual subscription and service terms in the proposed order rather than assuming a particular bundle or minimum engagement.

What is the first useful demonstration?

One permitted answer connected to its question, collection context, visible citations and exported file. It gives the buyer a concrete record to inspect before discussing broader totals.

What should we send Xindar?

Share the public website, the offer being evaluated, target markets, intended users and a few buyer questions. Identify the evidence fields and client-access rules that would make the pilot acceptable. Keep confidential customer material out of an initial public brief.

Sources and review boundary

The product explanation, public plans, platform overview, export-marketing service description and first-party research record were checked on October 31, 2026. The public dashboard was inspected separately on the same date.

Product-team statements, published descriptions and proposed procurement tests have different evidence status throughout this page. Recheck plan limits, prices, supported surfaces and contract conditions before purchase. The next substantive update should follow a changed official scope or a completed, permitted account test.