Open MCP-based standard for agentic advertising: media buy, creative, signals, and a measurement taxonomy. Complementary to adscapi on the conversion side.
AdCP (Ad Context Protocol) is an open, MCP-based standard for agentic advertising, launched in October 2025 by a 100+ company consortium including Scope3, Yahoo, and PubMatic, and governed by AgenticAdvertising.org. It standardizes the buy side — a Media Buy protocol (discover, buy, modify), a Creative protocol, and Signals — plus a measurement taxonomy of (metric_id, qualifier) tuples. It is a protocol/standard, not a library you import. adscapi is complementary on the measurement and conversion side: it interoperates with AdCP's measurement taxonomy and can emit a provide_performance_feedback request. It is not a media-buying agent.
| Dimension | adscapi | AdCP (Ad Context Protocol) |
|---|---|---|
| What it is | Code library you import and call | Open standard / protocol (MCP-based) |
| Scope | Conversion fan-out + measurement | Buy-side media, creative, and signals |
| How you use it | import { createAdscapi } from 'adscapi' | Implement an MCP agent against the spec |
| Conversion fan-out to 28 ad platforms | Yes — one track() to every configured CAPI | Not its job |
| Measurement | Can emit provide_performance_feedback; hashes and fans out conversions | Defines a (metric_id, qualifier) taxonomy agents implement |
| Relationship | Complementary — conversion / measurement side | Complementary — buy-side media and creative |
Billing is described by model only — check each vendor for current prices. adscapi is MIT-licensed open source; you bring your own platform tokens.
adscapi is free and open source. Set your platform tokens, call track() once, and every configured Conversions API receives the event — hashed the way each platform expects.
Back to the compare hub for every competitor, or read the docs to install.