Agent Development Kit
Description under review
Key facts
- Company
- Google Cloud
- Stack layer
- Agents
- Sector
- Agent Platforms
- Region
- Global
- Users
- Not disclosed
- Pricing
- api — Usage-based or enterprise pricing
- API
- Not published
What it does
A product-specific description has not been verified yet. Follow the official product link and check availability, capabilities and pricing before relying on this entry.
Decision support
Who it fits—and what to verify
Teams building or operating multi-step AI agents
- — No public API documentation is recorded
- — No product-specific adoption evidence is recorded
- — The catalog assessment does not yet include source URLs
- 01Which actions require human approval?
- 02Can runs be replayed, evaluated, and audited?
- 03How are permissions, secrets, and tool failures isolated?
This baseline is generated from recorded catalog fields; verify it against the linked evidence. Evidence level: partial.
Product Strength
One score, three components · 0–100
Unweighted geometric mean of Product Value, Adoption and Durability, each normalized to 0–1. All three count equally; displayed scores remain 0–100. Editorial estimate, not financial value.
Editorial range 0–64.5. 3/3 components rated. ≈ marks an editorial judgment, not a measured benchmark or a margin estimate. Open each component for its rationale and sources.
Comparison baseline: direct provider SDKs and a small application-owned orchestration loop · Limited source access; based on the recorded offering.
Product Value61
Expected usefulness for the intended job against named alternatives, including corrections and workflow completeness.
Agent Development Kit — Agent Development Kit: building and orchestrating tool-using agents. State, tools and evaluation integrations reduce plumbing, but add abstractions and operational complexity. The entry has limited descriptive material and no usable captured primary page; utility is assessed conservatively with wider uncertainty. Based on the existing catalog record and listed references; a usable primary-page capture was unavailable. Current capability and commercial scope need additional verification.
Editorial range 34–88 · Editorial rating · Reviewed (UTC) 2026-10-22
Adoption25
Individual products: active use. Enterprise products: product revenue or ARR first, conservative deployment evidence when undisclosed. A score, not market share. Parent scale, funding and downloads do not count.
Agent Development Kit — Agent Development Kit: building and orchestrating tool-using agents. The record identifies an offered product but does not establish its active-user or production-deployment scale. This is a conservative public-traction assessment, not a claim that actual adoption is small. The entry has limited descriptive material and no usable captured primary page; utility is assessed conservatively with wider uncertainty. Based on the existing catalog record and listed references; a usable primary-page capture was unavailable. Current capability and commercial scope need additional verification. Adoption uses diminishing returns: established use earns most of the credit, and additional audience scale earns progressively less. This is a score, not market share or retention.
Editorial range 0–50 · Editorial rating · Reviewed (UTC) 2026-10-22
Durability34
Ability to maintain and renew an advantage over three years against credible rivals: their strength and reach, bundling pressure, defensible assets and demonstrated expansion. Company size or competitor count alone earns no points.
Agent Development Kit — Agent Development Kit: building and orchestrating tool-using agents. Developer integrations and trained teams provide some stickiness; an open framework cannot rely on proprietary model access. The entry has limited descriptive material and no usable captured primary page; utility is assessed conservatively with wider uncertainty. Based on the existing catalog record and listed references; a usable primary-page capture was unavailable. Current capability and commercial scope need additional verification.
Editorial range 7–61 · Editorial rating · Reviewed (UTC) 2026-10-22
- Global reach · descriptive
- Not established
- AI role · descriptive
- Not established
Calculated 2026-10-27 · Rubric, evidence and limits
Pricing
Where it sits in the AI stack
View all in Agent Platforms (61)L5 · Vertical
Industry applications, robots, autonomous vehicles and task-specific physical systems
L4 · Agents
AI agent frameworks, orchestration, workflow automation
L3 · Infra
Inference, MLOps, developer tools, coding assistants
L2 · Models
LLMs, multimodal models, sovereign AI
L1 · Hardware
Reusable compute and memory: GPUs, NPUs, LPUs, AI servers and HBM
Company
View company profile →Google Cloud provides enterprise model and agent infrastructure. Gemini Enterprise Agent Platform succeeds Vertex AI; Model Garden, Agent Studio and Agent Garden belong to that platform. The employee-facing Gemini Enterprise app is a separate offering.
- Stage
- Public
- Category
- Developer and AI Infrastructure
- Headquarters
- Global
- Founded
- 2008
- Total funding
- Public
- Founder
- Not disclosed · Alphabet division
Competitors
Similar products
Same-type sector peers by estimated Strength
More from Google Cloud
Other products from this company
Google Cloud
Gemini Enterprise Agent Platform, the business agent and Code Assist are covered in the shared Gemini family. Google Cloud provides these enterprise workflows.
Image and video understanding APIs
Video metadata and content intelligence APIs
Text understanding API for classification and sentiment
Speech recognition API
Voice synthesis API
Machine translation API
Conversational AI and agent-assist for contact centers
Commerce AI product for discovery, merchandising, marketing, pricing, support, or retail operations.
Physical AI, robotics, chip, or autonomous-system product where model capability is coupled to hardware, sensors, or operational environments.
Physical AI, robotics, chip, or autonomous-system product where model capability is coupled to hardware, sensors, or operational environments.
Physical AI, robotics, chip, or autonomous-system product where model capability is coupled to hardware, sensors, or operational environments.
Google Cloud's fifth-generation TPU accelerators
Evidence & sources
Sources
Recorded
Adoption
Missing
Rating rationale
Recorded
Fact review date
Missing
Why this score
The entry has limited descriptive material and no usable captured primary page; utility is assessed conservatively with wider uncertainty.
Sources
Freshness & history
Material changes are dated; missing dates remain explicit.
No material rating action is recorded for this entity.
Coverage gaps
No source URLs · No product-specific adoption