StackAdapt MCP: report on StackAdapt in 
Claude & 
ChatGPTPorter Metrics connects StackAdapt to Claude, ChatGPT and other AI clients through one MCP server built for programmatic reporting. MCP is the Model Context Protocol, the open standard AI apps use to reach other tools. Ask which channel converts cheapest, which campaign group spends without results or what each client paid in fees, and the answer comes from live StackAdapt data.
Every metric and dimension Porter reads is in the StackAdapt fields list.
One server, every advertiser: Porter connects StackAdapt to the AI clients, fields and workflows programmatic teams already use. The same Porter server URL works in Claude, ChatGPT and any MCP client that supports remote servers. Porter reads StackAdapt with the API key you add once, and you pick the advertisers it reaches.
Works in: Claude (Claude.ai, Claude Desktop, Cowork and Claude Code), ChatGPT, Cursor and any MCP client that supports remote servers.
Account objects: advertisers · campaign groups (formerly line items) · campaigns · ads
Campaign setup: Advertiser · Campaign Group · Campaign · Campaign Channel · Campaign State · Campaign Status · Campaign Timezone
Creative: Ad · Ad Channel · Ad State · Ad Status · Creative Size
Delivery: Cost · Impressions · Unique Impressions · Frequency · eCPM · Clicks · CTR · eCPC
Conversion: Conversions · Click Conversions · Impression Conversions · Unique Conversions · Conversion Revenue · eCPA · ROAS · CVR
Video and audio: Video Starts · Video Q1 Playbacks · Video Q2 Playbacks · Video Q3 Playbacks · Video Completions · Video Completion Rate · eCPV · Audio Starts · Audio Completions · Audio Completion Rate
Viewability and engagement: Measured Impressions · Viewed Measured Impressions · Engagements · Engagement Rate · ATOS
Fees and margins: CPD Media Cost · Ad Serving · Targeting · Viewability Fee · Enterprise Features · Revenue · Profit · Margins · rCPM
Site and app: GA4 Sessions · GA4 Engaged Sessions · GA4 Bounce Rate · App Install · MMP CPI
Porter has 98 ready-made StackAdapt fields, 80 metrics and 18 dimensions. Each one is in the full StackAdapt field reference.
Hierarchy: Advertiser · Campaign Group · Campaign · Ad
Channel and creative: Campaign Channel · Ad Channel · Creative Size
Status: Campaign State · Campaign Status · Ad State · Ad Status
Time: Date · Campaign Timezone · Currency
Send work from Claude or ChatGPT to: email on a schedule
Turn answers into: live dashboards · client reports · weekly fee reviews
Same data in other AI tools: StackAdapt connector for ChatGPT · all MCP connectors
StackAdapt labels each daily bucket in UTC, and every amount comes in the advertiser’s own Currency. Porter splits results by advertiser, campaign group, campaign, ad, channel and day. Audience, geo, device, supply source and hourly splits stay in StackAdapt’s own insights reports.
Porter’s StackAdapt MCP reads your advertisers through an API key and changes nothing in StackAdapt. Claude reads campaigns, fees and results while budgets, bids and targeting stay exactly as you set them. You pick the advertisers Claude sees, and you can remove the key from Porter whenever you like.
Every StackAdapt question is a report. New campaigns, budget changes and pauses stay in StackAdapt, where your team handles them as usual.
StackAdapt issues API keys through its team rather than a self-serve page, per StackAdapt’s API docs. You paste the key in Porter once, and Claude talks to Porter instead of StackAdapt.
One key can reach many advertisers. You tick the ones Claude should read, and each one becomes its own Porter account.
Reporting is billed per connected StackAdapt advertiser. One more question about the same advertiser adds nothing to the bill.
Remove the key in Porter, or ask your StackAdapt account team to rotate it, and Porter’s access ends. Add a new key later to connect again.
Porter’s StackAdapt MCP uses one server URL in every AI client, plus a StackAdapt API key you add once. Claude takes Porter as a plugin or a custom connector. ChatGPT takes it through the Porter Metrics app or developer mode. Both read the same StackAdapt advertisers, with the same 98 fields.
Use the Claude plugin for the guided setup. Claude keeps the Porter plugin synced when new tools ship.
portermetrics/porter.portermetrics/porter.Anthropic’s guide to custom connectors using remote MCP walks through the same path, and custom connectors work on Free plans with a limit of one. The Claude-specific setup and use cases live on the StackAdapt connector for Claude page.
Click-by-click Claude plugin demo:
Use manual setup when Claude asks for a custom connector instead of a plugin.
Porter.https://mcp.portermetrics.com/mcp.StackAdapt to Claude demo:
Use the Porter Metrics app for the fastest ChatGPT route.
OpenAI documents connected apps in ChatGPT and custom MCP servers in ChatGPT developer mode. The ChatGPT side of the same connection is covered in full on the StackAdapt connector for ChatGPT page.
Click-by-click StackAdapt to ChatGPT demo:
Use developer mode when you want the full Porter setup in ChatGPT: connector plus skill.
Porter Metrics.https://mcp.portermetrics.com/mcp as the MCP Server URL.https://mcp.portermetrics.com/skill.zip.skill.zip to your ChatGPT project.StackAdapt hosts its own server, the StackAdapt MCP Server, at one URL:
StackAdapt’s guide to connecting its MCP server adds it to Claude as a custom connector and to ChatGPT in developer mode, with a StackAdapt sign-in. Use Porter’s URL when StackAdapt is one channel of several and you want fees, dashboards and scheduled emails in the same place.
Try this one once you’re connected:
Each StackAdapt skill in Porter’s MCP is a ready-made prompt for one question programmatic buyers ask every week. A skill turns live StackAdapt data into a channel ranking, a fee review, a cross-channel answer or a scheduled email. Copy a prompt into Claude or ChatGPT, then change the dates, advertisers or thresholds to fit yours.
The research skills blend StackAdapt with your ad and CRM accounts, so connect Google Ads, Meta Ads or HubSpot in Porter first.
These run on a schedule and send what they find, while your StackAdapt campaigns stay as they are.
Teams that buy programmatic for several clients or channels get the most from Porter’s MCP. Agencies, programmatic buyers, B2B demand-gen marketers and video and CTV buyers use it to turn a StackAdapt question into a channel ranking, a fee review or a cross-channel answer in one chat.
Marketers who report on StackAdapt next to the channels it competes with for budget.
Your client isn’t asking about eCPM. They’re asking what programmatic cost them per conversion, fees included.
You run StackAdapt only and want the results its MCP server pulls from internal endpoints, such as audience-level answers. StackAdapt’s server covers that job, and you can run it next to Porter.
StackAdapt’s own StackAdapt MCP Server and Porter’s MCP both bring StackAdapt data into AI clients. StackAdapt’s MCP server docs describe read tasks only in the initial release, using curated internal endpoints. Porter covers StackAdapt reporting and adds every other channel, fee reviews, dashboards and scheduled emails.
StackAdapt’s MCP hands your AI StackAdapt’s own view of StackAdapt. Porter’s MCP hands it StackAdapt next to everything else you run.
Both are hosted, use one URL and change nothing in your campaigns. Here is where they part.
| Porter | StackAdapt MCP Server | |
|---|---|---|
| Same job, different reach | ||
| Server | Hosted by Porter at mcp.portermetrics.com/mcp | Hosted by StackAdapt at mcp.stackadapt.com |
| Sign-in | An API key you add in Porter once | Each user signs in to StackAdapt |
| Data | 98 named fields by advertiser, campaign group, campaign, ad, channel and day | Curated tools, including internal endpoints not in StackAdapt’s public API |
| Changes to campaigns | None; reports only | None in the initial release; read tasks only |
| Claude plan | Every plan, Free included, as a custom connector | Free (one custom connector), Pro, Max, Team and Enterprise |
| Price | Included in every Porter plan, the free plan too | No added cost, per StackAdapt’s launch release |
| Only one side | ||
| Google Ads, Meta, HubSpot and GA4 in the same answer | Yes, from Porter’s 30+ data sources | — |
| Fee and margin review across advertisers | Yes | — |
| Live dashboards at their own link | Yes | — |
| Scheduled reports by email | Yes | — |
| Skills: whole StackAdapt tasks, already written | Yes | — |
StackAdapt’s column follows StackAdapt’s MCP server docs, its connect guide and StackAdapt’s MCP Server launch release (as of November 2026).
Use StackAdapt’s own if StackAdapt is your only channel and you want answers from its internal endpoints. Everything under Only one side is what happens when StackAdapt is one channel of several. Both connectors can sit in the same Claude account.
StackAdapt’s public API is for developers who build their own tools. Its MCP server and Porter’s MCP are for people who want answers in an AI chat.
A few more turn up for this search, and they split two ways.
Porter reads StackAdapt next to your marketing from one hosted server, starting on the free plan.
Every Porter plan includes the StackAdapt MCP at no extra cost, the free-forever plan too. Porter bills Reporting per connected data source account, and each StackAdapt advertiser counts as one. StackAdapt questions are reads, so they use no automation credits. Full details are on Porter pricing.
| What you use | What you pay | What is included |
|---|---|---|
| StackAdapt MCP | Included in every plan | The server URL, Claude and ChatGPT setup, reports, skills and scheduled emails |
| Reporting | From $12.50/mo per data source account, billed annually | Reports and dashboards across StackAdapt and your other sources. The free plan keeps unlimited connected accounts, up to 30 days of history and 15 days in AI chats |
| Automation | By credits | Approved changes on your other sources, such as pausing a Meta ad. The free plan includes 100 credits a month |
Free means usable: the free plan connects unlimited StackAdapt advertisers, keeps up to 30 days of history and gives you 100 credits a month.
Yes. StackAdapt announced general availability of its MCP Server on April 21, 2026, at https://mcp.stackadapt.com, with read tasks only in its first release. Porter’s StackAdapt MCP runs at https://mcp.portermetrics.com/mcp and adds your other channels, fee reviews and dashboards. You can connect both in Claude.
Porter’s StackAdapt MCP server URL is https://mcp.portermetrics.com/mcp. StackAdapt’s own server uses https://mcp.stackadapt.com. Paste either into Claude, ChatGPT or another AI client as a custom connector. Porter’s URL also reaches your Google Ads, Meta and HubSpot accounts.
Porter’s StackAdapt MCP is included in every Porter plan, the free-forever plan too. The free plan covers unlimited connected accounts, up to 30 days of history, 15 days in AI chats and 100 automation credits a month. Paid Reporting starts at $12.50 a month per data source account, billed annually, and the 14-day trial needs no card.
Porter reports on advertisers, campaign groups, campaigns and ads, and you pause, edit or re-budget them in StackAdapt as usual. That keeps your live campaigns exactly as you set them. Claude still tells you what to change, such as which channel to shift budget toward, with the numbers behind it.
Yes. One API key reaches every advertiser it can see, and you pick the ones Claude should read. Claude compares them in one answer with Advertiser as a column. Reporting is billed per connected advertiser, and the free plan connects unlimited accounts.
StackAdapt doesn’t offer a self-serve key page, so ask your StackAdapt account team for one. Paste the key in Porter, pick the advertisers and Claude starts reading them. If the key is rotated, paste the new one in Porter and your questions keep working.
Install the Porter plugin from Claude Settings → Customize → Plugins. Add the portermetrics/porter marketplace, install Porter Metrics from the Directory and add your StackAdapt API key in Porter when the connector asks. Then ask Claude for your StackAdapt advertisers.
MCP is the Model Context Protocol, an open-source standard for connecting AI applications to external systems like StackAdapt. MCP isn’t the agent. Claude and ChatGPT are the agents, and Porter’s MCP is how they reach your StackAdapt advertisers.
Check the day boundaries first: StackAdapt labels each daily bucket in UTC, while a campaign can run in its own Campaign Timezone, so compare by week or month. Amounts stay in each advertiser’s Currency. Campaign State and Campaign Status filter differently, and conversions from every tracker are summed.
Re-enter the API key in Porter and select the advertiser. Claude shows nothing when the key was rotated or revoked, or the advertiser isn’t selected in Porter. On the free plan, ranges past the history limit also come back empty, so ask for the last 30 days first.
Porter reads StackAdapt by advertiser, campaign group, campaign, ad, channel and day. Audience, geo, device, supply source and hourly splits live in StackAdapt’s insights reports. Ask Claude at the levels Porter reads, and it ranks and compares them for you.
Connect StackAdapt to Claude or ChatGPT and start with channel rankings, fee reviews and dashboards that land in your inbox.
Start free: unlimited connected accounts on the free-forever plan, or a 14-day trial with everything unlimited and no card.