Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
23 commits
Select commit Hold shift + click to select a range
7169038
Update MCP connector docs to match the current servers
ErikBernskiold Oct 23, 2026
e501229
Fix outdated REST API docs
ErikBernskiold Oct 23, 2026
83b2951
Document new REST API endpoints
ErikBernskiold Oct 23, 2026
06b7d0c
Remove Embed from the navigation and homepage
ErikBernskiold Oct 23, 2026
27e0573
Move the data models under Brand Data
ErikBernskiold Oct 23, 2026
75f218b
Document authentication with applications, API keys and OAuth 2
ErikBernskiold Oct 23, 2026
66e5ea1
Backfill blog posts for API and MCP releases since November 2025
ErikBernskiold Oct 23, 2026
c476a28
Recommend OAuth over long-lived API keys
ErikBernskiold Oct 24, 2026
cc801de
Remove references to the legacy auto-login scope
ErikBernskiold Oct 24, 2026
2517c13
Describe application access without the underlying user
ErikBernskiold Oct 24, 2026
4ec99d8
Remove the legacy X-User-Email header from the docs
ErikBernskiold Oct 24, 2026
befc79d
Remove the note about creating unknown users
ErikBernskiold Oct 24, 2026
e3da7ce
Split the tools out of Brand Data into their own section
ErikBernskiold Oct 24, 2026
9f4411a
Rewrite getting started, customizing and embed docs in Simplified Tec…
ErikBernskiold Oct 25, 2026
2ad34bd
Rewrite brand data and tools docs in Simplified Technical English
ErikBernskiold Oct 25, 2026
9e55502
Rewrite audiences, Audience Origin, Best Countries and brands docs in…
ErikBernskiold Oct 25, 2026
4646b60
Rewrite company, market share, measures, metadata, metrics and studie…
ErikBernskiold Oct 25, 2026
b5979aa
Rewrite AI and MCP docs in Simplified Technical English
ErikBernskiold Oct 25, 2026
738c1e4
Rewrite Python SDK docs in Simplified Technical English
ErikBernskiold Oct 25, 2026
6473d72
Fix Python SDK docs to match the bavapi source
ErikBernskiold Oct 25, 2026
f329f80
Fix errors and inconsistencies in the REST and MCP docs
ErikBernskiold Oct 25, 2026
94a9914
Use bavsupport@wpp.com, and remove unsent webhook events and unavaila…
ErikBernskiold Oct 26, 2026
d3a7168
Sync package-lock.json with package.json so npm ci passes
ErikBernskiold Oct 26, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion blog/2023-02-12-version-2-0.md
Original file line number Diff line number Diff line change
Expand Up @@ -23,4 +23,4 @@ At a high-level, these are the changes that will affect you:
- Ability to scope API keys to specific operations.

If you are having problems with using the API, or if you have any questions or feedback,
please [do reach out to us](mailto:support@bernskiold.com). We are happy to help.
please [do reach out to us](mailto:bavsupport@wpp.com). We are happy to help.
2 changes: 1 addition & 1 deletion blog/2023-04-20-upgraded-brandscape-data.md
Original file line number Diff line number Diff line change
Expand Up @@ -34,4 +34,4 @@ GET /api/v2/brandscape-data/?filter[sectors]=1
You can read more in the [Brandscape Data endpoint docs](/docs/2.x/brand-data/brandscape-data).

If you are having problems with using the API, or if you have any questions or feedback,
please [do reach out to us](mailto:support@bernskiold.com). We are happy to help.
please [do reach out to us](mailto:bavsupport@wpp.com). We are happy to help.
2 changes: 1 addition & 1 deletion blog/2023-05-19-new-tool-endpoints.md
Original file line number Diff line number Diff line change
Expand Up @@ -15,4 +15,4 @@ The new endpoints are available for most of the TurboPitch tools. If one that yo
out to us, and we are happy to add it for you.

If you are having problems with using the API, or if you have any questions or feedback,
please [do reach out to us](mailto:support@bernskiold.com). We are happy to help.
please [do reach out to us](mailto:bavsupport@wpp.com). We are happy to help.
2 changes: 1 addition & 1 deletion blog/2023-08-15-new-cities-endpoint.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ For some of the data that we track, we need to map it to a city. For this reason
world with a population over 500 people, based on the [GeoNames](https://www.geonames.org/) database.

We have now made this database available through the API. You can find the documentation for the new endpoint in the
[API reference](/docs/2.x/core-resources/cities).
[API reference](/docs/2.x/metadata/cities).

The cities endpoint comes with a few filters that you can use to narrow down the results. For example, you can filter
by country or region, or for example retrieve capitals only.
6 changes: 3 additions & 3 deletions blog/2024-01-19-python-sdk-1-0.md
Original file line number Diff line number Diff line change
Expand Up @@ -20,10 +20,10 @@ If you encounter any bugs while using the SDK, or if you have any suggestions/re

Some highlights since our last announcement:

- :tada: Full support for many more [endpoints](../docs/2.x/developer-tools/python/endpoints), and more [versatility](../docs/2.x/developer-tools/python/basic-usage#using-query-objects) for performing endpoint requests.
- :tada: Full support for many more [endpoints](/docs/2.x/developer-tools/python/endpoints), and more [versatility](/docs/2.x/developer-tools/python/basic-usage#using-query-objects) for performing endpoint requests.
- :rocket: `bavapi` queries should now be faster and more reliable, thanks to various improvements in the requests and response parsing logic.
- :sparkles: New [error handling](../docs/2.x/developer-tools/python/basic-usage#error-handling) behavior. If a page fails to download, `bavapi` will now warn the user about the missing pages instead of raising an exception. You can change this by using `on_errors="raise"` in endpoint calls.
- :notebook: Improved [SDK documentation](../docs/2.x/developer-tools/python) and updated with the latest features and functionality.
- :sparkles: New [error handling](/docs/2.x/developer-tools/python/basic-usage#error-handling) behavior. If a page fails to download, `bavapi` will now warn the user about the missing pages instead of raising an exception. You can change this by using `on_errors="raise"` in endpoint calls.
- :notebook: Improved [SDK documentation](/docs/2.x/developer-tools/python) and updated with the latest features and functionality.
- :bug: Fixed various issues with validation, data parsing and pagination.

Read the detailed changelog on [GitHub](https://github.com/wppbav/bavapi-sdk-python/releases).
2 changes: 1 addition & 1 deletion blog/2025-05-28-mcp-server.md
Original file line number Diff line number Diff line change
Expand Up @@ -13,4 +13,4 @@ Our new remote MCP server uses the latest MCP (Model Control Protocol) standard.
We're excited to see how you and your agents use BAV data to power your workflows. If you have questions, feedback, or
requests for new MCP tools, please let us know.

[Read more about the MCP Server](/docs/2.x/ai/mcp-server)
[Read more about the MCP Server](/docs/2.x/ai/mcp/intro)
24 changes: 24 additions & 0 deletions blog/2026-02-14-new-bav-mcp-data-tools.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
---
slug: new-bav-mcp-data-tools
title: Six new data tools on the BAV MCP server
authors: erikbernskiold
---

The BAV MCP server can now answer a much wider range of brand questions. We've added six new data tools, bringing
more of the analyses you know from the platform to your agents.

<!-- truncate -->

- **Commitment Funnel** (`get-commitment-funnel-data`) shows how consumers move from awareness to preference and
loyalty.
- **Perception SWOT** (`get-perception-swot-data`) highlights a brand's strengths, weaknesses, opportunities and
threats.
- **Category Drivers** (`get-category-drivers-data`) shows which attributes drive preference in a category.
- **Archetypes** (`get-archetypes-data`) maps brands to their archetypes.
- **Partnership Exchange** (`get-partnership-exchange-data`) finds brands with complementary imagery for partnerships.
- **Toplist** (`get-toplist-data`) ranks the top brands in a market on a given metric.

Agents can also create brand collections with the new `create-collection` tool, making it easy to save a set of
brands for later analysis.

See the [BAV MCP Server](/docs/2.x/ai/mcp/bav) documentation for the full list of tools and their parameters.
16 changes: 16 additions & 0 deletions blog/2026-02-27-search-result-limit.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
---
slug: search-result-limit
title: Search results are now limited to 1,000 records
authors: erikbernskiold
---

To keep searches fast and reliable, a search now returns at most the 1,000 best-matching records. This applies to
every API endpoint that supports the `search` parameter.

<!-- truncate -->

Most searches return far fewer results than this, so you are unlikely to notice a difference. If you need to go through
a very large set of records, combine filters rather than a broad search term, or leave out the search and paginate
through the full list.

See [Searching](/docs/2.x/customizing/searching) for more details.
30 changes: 30 additions & 0 deletions blog/2026-03-15-mcp-apps-and-audiences.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
---
slug: mcp-apps-and-audiences
title: Interactive charts and audience building in MCP
authors: erikbernskiold
---

Agents connected to the BAV MCP server can now show interactive charts and build custom audiences, right in the
conversation.

<!-- truncate -->

## Interactive charts with MCP Apps

MCP Apps let a server return interactive interfaces alongside its data. In clients that support MCP Apps, the new
display tools render BAV charts directly in the chat:

- `display-powergrid-chart` for the PowerGrid
- `display-pillar-chart` for the brand pillars
- `display-imagery-chart` for imagery attributes
- `display-brand-relationship-chart` for brand relationships

Read more in [MCP Apps](/docs/2.x/ai/mcp/mcp-apps).

## Audience building

Agents can now create and manage custom audiences, add definitions and deploy them, using tools such as
`create-audience`, `add-audience-definition` and `deploy-audience`. Once deployed, an audience can be used across the
BAV data tools like any other.

See the [BAV MCP Server](/docs/2.x/ai/mcp/bav#audience-tools) documentation for the full list of audience tools.
23 changes: 23 additions & 0 deletions blog/2026-04-13-correspondence-maps-and-media-performance.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
---
slug: correspondence-maps-and-media-performance
title: Correspondence maps and media performance in MCP
authors: erikbernskiold
---

Two new capabilities have arrived on the BAV MCP server: correspondence maps and media performance benchmarks.

<!-- truncate -->

## Correspondence maps

The new `get-correspondence-map-data` tool runs a correspondence analysis over imagery attributes and returns the
coordinates of brands and attributes on a shared map. It's the same analysis that powers the Correspondence Map in the
platform, and is great for finding whitespace, ownership of attributes and clusters of similar brands.

## Media performance

For brands that use media performance benchmarking, a new set of tools gives agents access to campaign performance
data and benchmarks, such as `get-campaign-performance-data` and `get-campaign-performance-benchmarks`. Access is
granted per brand, so the tools only appear for users who have access to at least one brand.

See the [BAV MCP Server](/docs/2.x/ai/mcp/bav) documentation for details.
18 changes: 18 additions & 0 deletions blog/2026-05-11-brandscape-data-on-demand.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
---
slug: brandscape-data-on-demand
title: Brandscape data is now generated on request
authors: erikbernskiold
---

The [Brandscape Data](/docs/2.x/brand-data/brandscape-data) endpoint now makes sure the data you ask for exists. If
a combination of study and audience hasn't been generated yet, requesting it starts the generation for you.

<!-- truncate -->

While data is being generated the endpoint responds with **202 Accepted** instead of 200 OK, and includes a
`meta.generation` block. Its `status` is `pending` when nothing is ready yet, or `partial` when some of the requested
data is available and the rest is on its way. Retry the request after a short while to get the complete data. Once
everything is available, you'll get a 200 OK as usual.

We've also added a safeguard for very wide requests. If you don't filter by study, country or year, the endpoint
now returns data for the latest available study rather than scanning all of BAV.
18 changes: 18 additions & 0 deletions blog/2026-05-21-agency-change-predictor-api.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
---
slug: agency-change-predictor-api
title: Agency Change Predictor in the API
authors: erikbernskiold
---

The Agency Change Predictor is now available in the API. It runs ten weighted checks
on a brand's equity, compared with its category, to assess how likely the brand is to change agency.

<!-- truncate -->

Send a request to `/api/v2/tools/agency-change-predictor` with the brands and studies (or countries and years) you're
interested in, and you'll get a prediction for each brand.

Since launch, the endpoint has been brought in line with the other data model endpoints. It accepts arrays of IDs
or names, supports study types, and returns the standard `meta` and `data` response.

See [Agency Change Predictor](/docs/2.x/brand-data/agency-change-predictor) for the full documentation.
21 changes: 21 additions & 0 deletions blog/2026-06-17-mcp-metrics-and-sharing.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
---
slug: mcp-metrics-and-sharing
title: Focused data and sharing collections in MCP
authors: erikbernskiold
---

We've made several improvements to the BAV MCP server to help agents get exactly the data they need, and share their
work.

<!-- truncate -->

- **Metric subsets**: The imagery, personality, PowerGrid, preferences, usage, recommendation, influence and pillar
tools accept a new `metrics` parameter. Agents can ask for just the metrics they need, which keeps responses smaller
and more focused.
- **Clearer metric keys**: Metric keys now say what type of data they hold, so models are less likely to mix up ranks
and percentages.
- **Sharing collections**: The new `share-collection` tool shares a brand collection with up to 10 people by email, with
view or edit access.
- **Creative Premium**: The new `get-creative-premium-data` tool is available to users with access to Creative Premium.

See the [BAV MCP Server](/docs/2.x/ai/mcp/bav) documentation for details.
15 changes: 15 additions & 0 deletions blog/2026-07-31-rank-or-percentage-in-mcp.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
---
slug: rank-or-percentage-in-mcp
title: Choose between ranks and percentages in MCP
authors: erikbernskiold
---

The imagery, personality, preferences, recommendation and usage tools on the BAV MCP server now take an optional
`score_type` parameter. Set it to `rank` or `percentage` to choose how the data is returned.

<!-- truncate -->

If you leave it out, each tool keeps its existing default, so nothing changes for current agents. Imagery and
personality default to percentile ranks, while preferences, recommendation and usage default to percentages.

Not sure which to use? Our new [Concepts for Agents](/docs/2.x/ai/mcp/concepts) page explains the difference.
20 changes: 20 additions & 0 deletions blog/2026-08-10-aura-in-mcp.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
---
slug: aura-in-mcp
title: AURA AI brand research in MCP
authors: erikbernskiold
---

AURA, our AI brand research, is now available on the BAV MCP server. AURA studies measure how AI models perceive brands,
using the same models as our consumer research.

<!-- truncate -->

Seven new tools give agents access to AURA data:

- `get-aura-studies` and `get-aura-audiences` to find the available studies and audiences.
- `get-aura-pillar-data`, `get-aura-powergrid-data`, `get-aura-personality-data` and `get-aura-imagery-data` for the
core brand models.
- `get-aura-drift-data` to compare how AI sees a brand against how consumers see it.

The tools are available to users with access to AURA. See the
[BAV MCP Server](/docs/2.x/ai/mcp/bav#aura-tools) documentation for details.
34 changes: 34 additions & 0 deletions blog/2026-08-20-explicit-study-types.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,34 @@
---
slug: explicit-study-types
title: Explicit study types across the API and MCP
authors: erikbernskiold
---

BAV data comes from different types of studies: our standard studies, SnapBAV, AO+BAV Fused and custom studies. They
aren't interchangeable, so you now always choose which type of study your data comes from.

<!-- truncate -->

## In the API

The brandscape endpoints accept a new `filter[study_types]` parameter, with the values `syndicated`, `snap`, `fused`
and `custom`. It defaults to `syndicated`. If a market is only covered by another study type, you get no rows and a
`meta.available_alternatives` block that tells you what exists and how to ask for it. Each row now also includes its
study type.

## In MCP

The brandscape tools now work from studies. If an agent asks for a country and year, it gets a list of matching studies
with their type, caveats and a recommendation, and is asked to pick one. Models handle this well, but it is a change in
behaviour worth knowing about. The `get-markets` tool now returns each study's ID as `study_id`.

## Other changes

- The "BAV (Syndicated)" study type is now called **BAV (Standard)**. The `syndicated` value in the API and MCP stays
the same.
- Audiences now include population projections. `GET /api/v2/audiences/{id}/studies` returns a `population` per study,
and the new `get-audience-studies` MCP tool does the same.
- The audience tools on the BAV MCP server are now available to all users.

Read more in [Brandscape Data](/docs/2.x/brand-data/brandscape-data) and
[Concepts for Agents](/docs/2.x/ai/mcp/concepts).
22 changes: 22 additions & 0 deletions blog/2026-09-08-health-check-and-acp-updates.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
---
slug: health-check-and-acp-updates
title: Health check without a token, and Agency Change Predictor updates
authors: erikbernskiold
---

A couple of smaller improvements to the API.

<!-- truncate -->

## Health check without a token

The [health check](/docs/2.x/health-check) endpoint at `/api/v2/health` no longer needs an API token. That means you
can tell the difference between the API being down and your token having expired, and uptime monitors no longer need
credentials.

## Agency Change Predictor

The [Agency Change Predictor](/docs/2.x/brand-data/agency-change-predictor) now works like the other brandscape
model endpoints. It accepts arrays of IDs or names, honours study types and years, returns the standard `meta` and
`data` envelope, and includes each brand's classification. The comma-separated format still works, so existing
integrations keep running.
29 changes: 29 additions & 0 deletions blog/2026-09-12-market-share.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,29 @@
---
slug: market-share
title: Euromonitor market share in the API and MCP
authors: erikbernskiold
---

Market share data from Euromonitor is now available in both the API and through a new MCP server, alongside BAV
brand equity data.

<!-- truncate -->

## In the API

New endpoints give you Euromonitor's catalogs of brands, companies, categories, data types, geographies and industries,
and the brand and company share data itself. They need the `bav:market-share-data` scope, and access to Euromonitor
data on your account.

A few things to keep in mind:

- Euromonitor has its own IDs for everything. Fields that link to BAV are prefixed with `bav_`.
- A `null` value means the figure wasn't supplied, never zero.
- Share queries must be narrowed by at least one brand, company, category, geography or industry.

See [Market Share](/docs/2.x/market-share/) for the full documentation.

## In MCP

The new Market Share MCP server at `/v1/market-share` gives agents the same data. See the
[Market Share Server](/docs/2.x/ai/mcp/market-share) documentation for details.
24 changes: 24 additions & 0 deletions blog/2026-09-28-brand-tags-and-hierarchy.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
---
slug: brand-tags-and-hierarchy
title: Brand tags and brand hierarchy
authors: erikbernskiold
---

The brands endpoint can now tell you more about what a brand is, and how it relates to other brands and companies.

<!-- truncate -->

## Brand tags

Brand tags describe what a brand is, rather than how it performs. The new [Brand Tags](/docs/2.x/brands/brand-tags)
endpoint lists every tag, and you can filter brands with `filter[tags]` and include their tags with `include=tags`.
Tags can be named by their ID or their key.

In MCP, the `get-brands` tool has a new `tags` filter, and the new `list-brand-tags` tool lists the tags available.

## Brand hierarchy and companies

You can now filter brands by the company that owns them with `filter[companies]`. Each brand also includes its
`parent_brand_id`, and you can include its `parent` and `children` brands.

See [Brands](/docs/2.x/brands) for the full documentation.
18 changes: 18 additions & 0 deletions blog/2026-10-17-measures-and-questions.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
---
slug: measures-and-questions
title: Find which studies ask a measure or question
authors: erikbernskiold
---

New endpoints make it easier to find out what a study measures, and where a measure or question was asked, before you
query for data.

<!-- truncate -->

- `/bav-questions` lists the BAV audience questions, alongside the existing `/bav-question-groups`.
- `/brand-measures/{id}/studies` and `/bav-questions/{id}/studies` list the studies a measure or question was asked
in.
- `/studies/{id}/brand-measures`, `/studies/{id}/brand-metrics`, `/studies/{id}/audience-questions` and
`/studies/{id}/bav-questions` list what a study asked.

See [Measures & Questions](/docs/2.x/measures/) for the full documentation.
Loading
Loading