Skip to content

Framework

Google ADK

Google's open-source Agent Development Kit (ADK) is a Python/Java framework for building, orchestrating, and deploying multi-agent AI systems.

Known aliases

  • ADK
  • adk-python
  • Agent Development Kit
  • @google/adk
  • Google ADK
  • google.adk
  • Google ADK 2.6.2
  • Google ADK Java
  • google/adk-python
  • google.adk.workflow
  • Google's ADK

Current stories

buildConfirmed2 publishers

Workday ships Agent Harness as generally available, billed on Flex Credits

Workday made Sana Home the starting point for work and put Agent Harness into general availability, with both priced on Flex Credits. Agents built on it run in isolated environments over finance and HR data, Constellation Research reported.

Publishers:constellationr.com mezha.net

Reality

Evidence30
Adoption
Not scored
Hype gap+30
Incentives
Not scored
Confidence45
buildOne report1 publisher

Spotify's Ads AI ties each production agent to one package and one owning team

Spotify's Ads AI multi-agent platform has produced about 20,000 ad creatives for more than 7,000 advertisers, senior engineer Pratik Rasam said. Each agent is its own Bazel package naming the team that reviews its changes and answers its alerts, and other teams can copy that part most easily.

Publishers:infoq.com

Reality

Evidence45
Adoption55
Hype gap+20
Incentives65
Confidence55
buildOne report1 publisher

Regression-testing kagent agents in agentevals means ignoring ADK's role field

kagent-agentevals keeps 8 of 14 ADK events from a real kagent session when it builds an agentevals trajectory for regression tests. Its author found the role field crediting some agent tool calls to the user, so the converter reads part types and drops runtime adk_ tools.

Publishers:dev.to

Reality

Evidence45
Adoption
Not scored
Hype gap0
Incentives40
Confidence45
buildOne report1 publisher

Moving an agent definition into Postgres leaves seven lines of Python on disk

In apowerb the module the runtime imports is generated and disposable, and every instruction, tool and model string is read out of a database row at load time. The cost is a filesystem holding derived state.

Publishers:dev.to

Reality

Evidence60
Adoption
Not scored
Hype gap+10
Incentives80
Confidence55
buildOne report1 publisher

Wood Mackenzie consolidates three in-flight agent stacks onto one managed runtime

The company says 88 percent of its AI proofs-of-concept never reach widescale deployment, and it blames the identity, session and evaluation layers each team was rebuilding, so APEX builds them once on Amazon Bedrock AgentCore.

Publishers:aws.amazon.com

Reality

Evidence34
Adoption36
Hype gap+28
Incentives84
Confidence52
buildOne report1 publisher

Porting one Iceberg agent across three frameworks rewrites the invocation and auth code

A step-by-step build of the same Apache Iceberg data agent on Google ADK, AWS Strands and Microsoft Agent Framework measures what a cloud move costs in code, in latency and in storage wiring.

Publishers:dev.to

Reality

Evidence68
Adoption
Not scored
Hype gap0
Incentives35
Confidence55
buildOne report1 publisher

Deep Agents hands each subagent a fresh tool grant instead of narrowing the parent's

The framework never intersects a child's tool list with its caller's, so a supervisor holding one tool can spawn a writer that searches the web. The intersection is a middleware you have to write yourself.

Publishers:dev.to

Reality

Evidence68
Adoption20
Hype gap+12
Incentives78
Confidence62
buildOne report1 publisher

The National Archives catalog API answers a rate limit with 200 OK and a 5,454-byte HTML shell

The catalog serves its single-page app shell for malformed requests and for throttling alike, so the status line tells you nothing. The only gate that holds under load is checking the content type before you parse.

Publishers:dev.to

Reality

Evidence52
Adoption10
Hype gap−12
Incentives25
Confidence58
buildOne report1 publisher

Summarising ADK's event stream into four fields moves the assertion off the model's prose

A dev.to walkthrough puts deterministic tool tests at the base of a four-layer agent pyramid and saves live Gemini runs for the top. It works only if you own the adapter that turns framework events into a contract.

Publishers:dev.to

Reality

Evidence45
Adoption
Not scored
Hype gap+12
Incentives35
Confidence45
buildOne report1 publisher

A hackathon agent refuses to tick a form item until the question is actually answered

Shipping meeting assistants mark coverage the moment a topic comes up. Junwei Lai's intake adjudicates each open item in its own call and defaults to insufficient, which pushes the missing answer back into the room while the patient is still there.

Publishers:dev.to

Reality

Evidence42
Adoption6
Hype gap+14
Incentives55
Confidence52
buildOne report1 publisher

Giving one deterministic layer the only write privilege demotes the model to a proposer

An eldercare agent published at a live URL puts its two hard rules in Python instead of prompts, and the demo tests them by calling the send function directly, which is the only version of that claim a reader can check.

Publishers:dev.to

Reality

Evidence32
Adoption8
Hype gap+14
Incentives68
Confidence46
buildOne report1 publisher

File assignment alone moved the same model from 8.7% to 34.3% on one repository

A manager agent can only read the one dependency edge somebody wrote for a machine. That is why partitioning went back to a hand-maintained file, with a diff check standing between the engineers and the bookkeeping.

Publishers:dev.to

Reality

Evidence34
Adoption8
Hype gap+18
Incentives24
Confidence42
buildOne report1 publisher

CrisisMesh paged its floor warden about the same missing teacher every 25 seconds

One escalation loop got the wording right but sent it to the wrong person. No unit test could have caught that, because who gets paged and what stops the paging are both resolved outside the function under test.

Publishers:dev.to

Reality

Evidence46
Adoption8
Hype gap−18
Incentives58
Confidence47
buildOne report1 publisher

An ACT floor of 0.90 lands inside the model's own 0.85 to 0.95 confidence band

Sonjomon computes each action's autonomy from confidence and blast radius, which leaves a risk registry in code doing the load-bearing work while the confidence half rests on a number the model reports about itself.

Publishers:dev.to

Reality

Evidence44
Adoption8
Hype gap+14
Incentives48
Confidence41
buildOne report1 publisher

Splitting an ADK graph across A2A workers strands output_key in the worker's session

Google ADK's output_key writes into whichever agent's session declared it, and in-process that session is shared. A branch-coverage gate that only ever runs the single-process topology cannot reach the failing path.

Publishers:dev.to

Reality

Evidence48
Adoption18
Hype gap−12
Incentives58
Confidence46
buildOne report1 publisher

Cloud Run's agent card tells clients to reach the agent at 0.0.0.0:8080

A discovery-only comparison of A2A agents on three clouds, measured in August 2026, found one protocol producing three auth models and one card whose sole declared interface URL is the container's bind address.

Publishers:dev.to

Reality

Evidence58
Adoption
Not scored
Hype gap−5
Incentives32
Confidence55
buildOne report1 publisher

AgentCore charges a second IAM action just to read the agent card, and the card talks

A grant of InvokeAgentRuntime will not fetch an A2A agent card on Bedrock AgentCore. And the card a Strands agent publishes there runs 2,109 bytes, most of it the agent's own instructions.

Publishers:dev.to

Reality

Evidence62
Adoption18
Hype gap+8
Incentives38
Confidence55
buildOne report1 publisher

The A2A card said 0.0.0.0:8080, and the only caller it broke was Google's

One line of Google ADK turns an agent into an A2A server. Deployed to Cloud Run and called from AWS and Azure, it advertised the address it binds instead of the address anyone can dial.

Publishers:dev.to

Reality

Evidence58
Adoption18
Hype gap+8
Incentives32
Confidence54
buildOne report1 publisher

Three clouds, one protocol, no stored credentials: the runtime becomes a swappable part

A Bedrock AgentCore agent serving Google and Azure callers over A2A pushes the vendor-specific work into the container contract rather than the agent. The credential claim is the part not yet shown.

Publishers:dev.to

Reality

Evidence58
Adoption20
Hype gap+12
Incentives48
Confidence44

Earlier coverage