CopilotKit connects agents to product interfaces through AG-UI, chat and generative UI components, frontend actions and shared state. Developers can use ready-made components or headless APIs, with hosting and enterprise deployment options.
BackEngine collects customer interactions from connected business systems, matches them to accounts and provides governed context through MCP and other surfaces. Teams configure access and sensitive-data controls for account research and workflow preparation.
Progress Agent Engineering traces model calls, prompts, retrieval and tools, then connects costs and failures to evaluations and experiments. Native .NET, Python and TypeScript support helps teams investigate production agent behavior and review governance controls.
HAR coordinates coding workflows through a repository-owned harness contract, isolated worktrees and launch/verify/teardown stages. Its CLI and MCP expose project checks and evidence artifacts, while Mission Control tracks runs, validations and repositories.
Soloop supports planning, demand research, coding and growth workflows around a company goal. It can prepare recurring operating work and multi-channel activity, with pricing terms describing approval-first external actions and output review.
Remix gives teammates live sandbox variants of an existing codebase to edit with their AI coding tools. It records prompts, agent responses and file changes, then sends engineering a reviewable request with a preview link and configured guardrail results.
Omniwork coordinates specialized creative agents for planning, producing, revising and delivering content. It offers reusable workflows, team orchestration and project memory across social content, film, music and lightweight game projects.
dolv coordinates a Director agent and task-specific agents across marketing, outreach, content and CRM work. It drafts from company context, queues external actions for approval and associates connected analytics with campaigns and funnel goals.
Toolport connects MCP servers to multiple AI clients through one local gateway, manages reconnection and OS-keychain credentials, and flags changed tools. Optional sharing synchronizes server configuration while credentials remain per user.
AgentConnect brings multiple coding or ACP agents into team channels with roles, handoffs, schedules, tool access and shared visibility. A console manages sessions and permissions, while daemons run agents in selected execution environments.
Trigger.dev runs TypeScript tasks and durable AI agents with retries, queues, streaming and observability. Its chat-agent workflow supports paused tool approvals and continued execution, while deployments version task code and Realtime exposes run status.
Tines provides a governed workspace for AI agents, apps and business automations. Its 3B product brings visibility into generated work, centrally controlled access and model choices, and workflow changes that users review before deployment.
Statewave records immutable episodes, compiles typed memories and retrieves ranked context within a token budget. It adds source provenance, sensitivity labels, configured policies and integrity receipts, backed by self-hosted Postgres and optional connectors.
Sidekick is a Mac agent that responds to requests across files, browser pages and system tasks. A floating bubble or keyboard launcher accepts instructions, exposes action results and describes approval for sensitive operations and undo for supported changes.
Oqoqo runs repeated agent experiments on sandboxed cloud machines with chosen tasks, treatments and rubrics. It compares agents, models and tool packages, retaining trajectories, errors, files and pass/fail results for iterative interface improvements.
Kubit analyzes agent traces through execution-path clusters, Sankey flows, funnels and cohorts. It labels spans with its Polaris evaluator and joins agent behavior to business outcomes using traces in supported customer warehouses.
Gotcha accepts spoken or written requests and uses Android device context to inspect or perform system actions. Monitor mode plans with read-only tools; Operator mode enables permission-gated messages, app navigation, settings and screen automation.
BetterClaw configures hosted agents through chat and a visual workspace, connecting tools and messaging channels for scheduled tasks. It supports provider keys, per-agent access controls, execution history, sandboxing and a kill switch.
BearDrive synchronizes a chosen real folder across a team and its file-writing agents, with attributed version history and restorable changes. Files can remain team-gated or be shared through revocable public links, with read/freshness insights.
Ballet builds business playbooks from conversational requirements as reviewable, version-controlled code. Workflows combine agent reasoning with explicit logic, integrations, approval gates and run history across sales, support, finance and operations systems.
WebBrain is a browser extension that reads pages, extracts data and performs approved browser tasks with a chosen model. Read-only Ask mode is the default, while Act mode provides interactions; a local MCP bridge connects external coding clients.
Execlave checks governed agent actions against configured policies before execution and can allow, hold or block a request. It combines a registry, approval controls, spend limits and hash-chained audit evidence in managed or licensed self-hosted deployment.
Port22 attaches an iPhone to existing Claude Code, Codex or OpenCode terminal sessions on a Mac. Users read live transcripts, reply and approve actions remotely, with push notifications and local-network or relay connectivity.
Nuphos brings Claude Code and Codex into shared sessions with team tools, memory and infrastructure views. Users can switch between a local machine, their remote host and Nuphos Cloud, with schedules or webhooks starting follow-up work.
Not listed
API available
Checked Oct 9, 2026
Showing 168 of 1,345 tools
Plan facts carry individual sources and check dates. Unknown values are excluded from required filters.