Comparly.ai
Devs.ai

Devs.ai Review

4.7

Free (2 GB, core builders)

Visit Devs.ai

Governed agent building on 50+ models

Who it's for

Companies that need people outside engineering to build AI agents on internal data without creating shadow AI. The audit, permission and isolation model is the real centre of gravity here, not the builder itself.

Who it's not for

Solo builders and small teams who just want the cheapest route to a working agent. The per-seat plus per-usage pricing and the enterprise controls are overhead you will not use.

Pricing

PlanPriceNotes
Free$0/moAgent Builder and App Builder core features, 2 GB data upload. No workflows, APIs or integrations.
Personal$15/seat/moBilled annually. Includes $10 of usage per month, full builders, workflows, APIs and integrations, 10 GB.
Business$30/seat/moBilled annually. Includes $21 of usage per month, team governance and collaboration, 20 GB.
EnterpriseCustomAnnual contract. Adds SSO, ACLs, multi-tenancy and custom storage.
See the full Devs.ai pricing breakdown

Pros

  • Model-agnostic by design: OpenAI, Anthropic, Google, Meta, Cohere, Grok, Perplexity, DeepSeek and 50+ more, swappable per agent or per step without losing memory or workflows
  • Governance is built in rather than bolted on — isolated workspaces, RBAC, activity logs, audit trails, and contractual zero-training agreements with the LLM providers
  • RAG grounding from uploads, Google Drive, OneDrive, Jira, Slack, Confluence and Salesforce, plus any MCP server, with each source keeping its own access rules
  • Low/no-code studio lets non-engineers ship an agent, and a separate App Builder covers full internal apps rather than chat-only assistants

Cons

  • Seat-based pricing sits on top of metered usage — $15/seat/month only includes $10 of usage, so busy agents add cost on both axes
  • SSO, ACLs and multi-tenancy are Enterprise-only, which puts basic access control behind a custom annual contract
  • No documented multi-agent orchestration layer: agents are built and governed individually rather than composed into supervised teams

Alternatives