{"id":"dataelement-clawith","name":"Clawith","homepage":"https://www.clawith.ai","repo_url":"https://github.com/dataelement/Clawith","category":"ai-ml","subcategories":[],"tags":["ai-ml","multiagent","agents","collaboration","workflows","mcp","fastapi","react"],"what_it_does":"Clawith (OpenClaw for Teams) is an open-source multi-agent collaboration platform. It provides persistent agent identities, long-term memory and per-agent workspaces, plus an autonomous trigger system (cron/once/interval/poll/on_message/webhook) to coordinate agents. The platform is positioned as multi-tenant with RBAC, quotas, approval workflows, and audit logging, and it can integrate with chat/communication channels (e.g., Slack/Discord/Feishu/Lark) and with external tool ecosystems via an MCP client.","use_cases":["Team-based multi-agent workflows with persistent agents and long-term memory","Autonomous scheduling/monitoring via cron/interval/poll/webhook triggers","Organization-style collaboration: shared knowledge feed and auditability","Human-in-the-loop approvals for risky operations","Integrations with messaging platforms and external tool providers","Running a self-hosted agent “digital employee” system for internal operations"],"not_for":["Unmanaged single-user experimentation that doesn’t require tenancy/RBAC/auditability","Environments that cannot run self-hosted web applications (needs backend+frontend and supporting services)","Use cases requiring strict contractual uptime/SLA from a hosted SaaS vendor (this appears self-hosted)"],"best_when":"You want a self-hosted, team-oriented multi-agent system with persistence, scheduling/monitoring triggers, and org-style controls (RBAC, quotas, approvals, audit logs).","avoid_when":"You need a lightweight single-agent chatbot SDK only, or you cannot accept the operational overhead of maintaining a web app plus database (and optionally Redis).","alternatives":["CrewAI (self-hosted orchestration library)","LangGraph (agent workflows/state machines)","AutoGen (multi-agent conversations)","OpenAI Assistants / similar hosted agent platforms (if persistence/control are handled by the vendor)"],"af_score":43.2,"security_score":57.5,"reliability_score":32.5,"package_type":"skill","discovery_source":["openclaw"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-30T13:18:00.009500+00:00","interface":{"has_rest_api":true,"has_graphql":false,"has_grpc":false,"has_mcp_server":false,"mcp_server_url":null,"has_sdk":false,"sdk_languages":["Python","TypeScript"],"openapi_spec_url":null,"webhooks":true},"auth":{"methods":["JWT-based auth (JWT/RBAC mentioned)","RBAC with organization isolation"],"oauth":false,"scopes":false,"notes":"README indicates JWT/RBAC and organization-based isolation, but does not specify OAuth flows, scope granularity, or exact auth endpoints/headers."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"Appears self-hosted open-source; README discusses LLM inference via external providers (OpenAI/Anthropic), but does not provide pricing tiers for Clawith itself."},"requirements":{"requires_signup":true,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":43.2,"security_score":57.5,"reliability_score":32.5,"mcp_server_quality":20.0,"documentation_accuracy":50.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":65.0,"rate_limit_clarity":20.0,"tls_enforcement":70.0,"auth_strength":70.0,"scope_granularity":40.0,"dependency_hygiene":40.0,"secret_handling":60.0,"security_notes":"README recommends enabling HTTPS, setting strong SECRET_KEY/JWT_SECRET_KEY, changing default passwords, restricting Docker socket access, backing up regularly, and using PostgreSQL in production. However, the provided documentation does not include concrete implementation details (TLS enforcement defaults, secret handling mechanics, scope model details, or dependency/security posture), so scores reflect partial visibility from README only.","uptime_documented":0.0,"version_stability":50.0,"breaking_changes_history":40.0,"error_recovery":40.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["No explicit agent-facing API/MCP spec details provided in the README; integration may require inspecting backend code.","Multi-agent systems can produce complex state; if retrying trigger-fired actions, idempotency behavior is not described in README.","Trigger types include external HTTP monitoring and webhook receiving; correct security configuration (auth/signing, network exposure) must be handled carefully by the deployer."]}}