Categories
Explore the agent ecosystem by workflow category, release activity, and latest signal movement.
Categories
4
tracked workflow groups
Agents
37
across all categories
Updated 30d
15
active ecosystem signal
Category Map
Coding
AI agents that help write, review and debug code
Updated
13
Signals
373
Delta
25
Latest Signal
Cursor: Cursor launched the Agent SDK, now generally available, allowing developers to define agents as folders of code with instructions, tools, skills, MCP connections, and channels. Agents can be tested with evals locally or in CI and deployed to Cursor's infrastructure or self-hosted. Supported triggers include GitHub/GitLab/Bitbucket pull request events, Slack mentions, webhooks, and cron schedules.
Category Map
Research
AI agents that help find and synthesize information
Updated
6
Signals
125
Delta
18
Latest Signal
Cohere: Rerank 5 is a new retrieval optimization model featuring sharper search and fewer tokens. It also introduces a new feature called Context Focus. This is Cohere's latest Rerank model release.
Category Map
General
Versatile AI agents for a wide range of tasks
Updated
3
Signals
100
Delta
18
Latest Signal
Claude: Claude Haiku 5.5 is Anthropic's cheapest, fastest, and most capable small model, designed for high-volume, cost-sensitive tasks like summaries, database queries, and classification. It costs around 75% less than Haiku 4.5 to run and is the first Haiku-class model with an adjustable effort setting. Sonnet 5.5 cache read prices are also halved as part of this launch.
Category Map
Multimodal
AI agents that work with text, images, audio and video
Updated
2
Signals
58
Delta
12
Latest Signal
Gemini: Gemini 4 Argon introduces frontier-level cyberdefense capabilities designed to give defenders an advantage over attackers, with significantly improved performance on cybersecurity tests (solving all 12 end-to-end simulated cyberattack scenarios, up from five in the previous generation). The model features enhanced safety with violations reduced nearly twofold on adversarial stress tests, three layers of real-time protection against jailbreaks, and an automated misalignment monitor. It also includes improved boundary-respecting behavior for agentic tasks, with shortcut/cheating attempts dropping to just 2% on coding benchmarks.