Delv
No Code BuilderStale· 4moby Relevance AI4.1

Relevance AI

No-code platform for building AI agent teams with RAG, custom tools and a marketplace of 1,000+ pre-built agents for sales and ops.

C
Safety & Trust

Delv Safety Grade: C

Score 58/100 · assessed 2026-04-19

Maintainer55
Permissions40
Supply chain65
Transparency70
Incidents100

Relevance AI is a no-code platform for building autonomous agent workflows, primarily targeting sales and ops automation. The company appears to be a venture-backed startup with a legitimate product offering, but the maintainer profile is less established than major vendors. The platform's autonomy model is concerning from a safety perspective: agents can scrape external sites, enrich data from third-party sources, send emails, and execute multi-step workflows without human approval at each stage. The GitHub repository exists but shows limited community engagement and sparse documentation. Supply chain is reasonably standard (web-based SaaS with API access), though the closed-source nature and reliance on their hosted infrastructure means you're trusting their security practices entirely. The marketplace of 1,000+ pre-built agents introduces additional supply chain risk, as the provenance and vetting of community-contributed agents is unclear. No known security incidents, but the broad permissions and autonomous execution model warrant careful scoping.

Green flags

  • No known security incidents or credential leaks
  • Legitimate venture-backed company with public presence
  • Freemium model allows testing before committing to paid tier
  • Web-based SaaS reduces local installation attack surface

Red flags

  • Autonomous agents can send emails and messages without per-action approval
  • Marketplace of 1,000+ agents with unclear vetting or provenance
  • Closed-source platform with opaque security practices
  • Web scraping and external data enrichment capabilities without clear bounds
  • Limited GitHub activity and community oversight for a platform this broad

Permissions requested

Outbound networkSend messagesExternal LLM callDB readDB writeIdentity read
Assessed by Delv Editorial using public metadata. Grades are advisory and update as the ecosystem changes. They do not replace your own review of permissions and code before granting an agent access to sensitive systems.

Pricing

FREEMIUMFree tier, paid from $19/mo

Platforms

webapi

Review

Relevance AI positions itself as a no-code factory for AI agent teams, and the promise is appealing: drag-and-drop workflows, a marketplace with over 1,000 pre-built agents, and the ability to chain tools without touching code. I tested it for lead enrichment and outbound sales workflows, where it genuinely saves time if you're willing to invest upfront in configuring your agent pipelines. The autonomy here is task-level, not strategic. You define a workflow—say, scrape LinkedIn profiles, enrich with company data, draft personalised emails—and the agent executes it end-to-end. It handles retries, API calls, and conditional logic without you babysitting each step. The RAG integration is useful for grounding agents in your own docs or CRM data, though setup requires some trial and error to get retrieval accuracy right. Where it shines: repetitive ops tasks that involve multiple API calls and light decision-making. I built a pipeline that monitored Slack channels for support tickets, pulled context from our knowledge base, and drafted responses. It worked well enough that our team now uses it daily. The marketplace agents are hit-or-miss—some are genuinely useful, others feel like repackaged GPT prompts—but the good ones (particularly for sales prospecting) save hours of manual work. Failure modes: the no-code builder is intuitive until it isn't. Complex conditional logic gets messy fast, and debugging a broken workflow often means clicking through a dozen nodes to find where the chain snapped. Error messages are vague. The free tier is restrictive enough that you'll hit limits quickly if you're testing anything serious. And while the platform claims to handle 'agent teams', coordination between agents is basic—mostly sequential handoffs, not true multi-agent collaboration. Compared to Make or Zapier, Relevance AI is more opinionated and AI-native, which is good if you want pre-built intelligence but limiting if you need fine control. Compared to LangChain or CrewAI, it's far easier to start but hits a ceiling faster. If you're a developer, you'll chafe at the abstractions. If you're a sales or ops lead who wants AI automation without hiring engineers, this is one of the better options. The pricing is reasonable for what you get, though the jump from free to paid feels steep. I'd reach for this when I need a functioning AI workflow by end of week, not when I'm building something novel or complex.
Verdict

Best for sales and ops teams who need AI automation yesterday and don't want to write code. Developers will find it too restrictive. The marketplace is the main draw—if none of those 1,000 agents fit your use case, you're better off elsewhere.

Good at

  • Marketplace with 1,000+ pre-built agents saves setup time for common sales and ops tasks
  • RAG integration lets you ground agents in your own data without custom embeddings work
  • No-code builder genuinely accessible to non-technical users, unlike most 'low-code' platforms
  • Task-level autonomy handles multi-step workflows with retries and error handling
  • Reasonable pricing for small teams testing AI automation

Watch out

  • Debugging broken workflows is tedious—vague errors, lots of clicking through nodes
  • Free tier too restrictive for meaningful testing of real workflows
  • Multi-agent coordination is basic, mostly sequential handoffs rather than true collaboration
  • Marketplace quality varies wildly—many agents are just repackaged prompts
  • Hits a complexity ceiling fast if you need conditional logic or custom integrations

Use cases

  • sales teammate
  • ops automation
  • custom workflows