Vellum

Build, evaluate, and deploy LLM apps and agents with confidence.

Hand-reviewed by the Toolglade team·Updated Aug 2026

coding#llmops#ai-agents#prompt-engineering#evaluation

Free plan Claimed API Self-hosted Teams

Toolglade’s take

Vellum is a solid all-in-one platform for teams that want to prompt, build, evaluate, and deploy LLM features together, especially with non-technical collaborators involved. Its evaluation tooling and visual workflow builder are real strengths. The main friction is opaque, sales-led pricing that makes budgeting hard, plus it is not open source and self-hosting is Enterprise-only. As a smaller company it also carries more platform-maturity risk than infrastructure incumbents.

About Vellum

Vellum is a commercial development platform for LLM apps and AI agents that combines prompt engineering, a visual workflow builder, evaluation and testing, and production deployment with monitoring. It supports many model providers and offers enterprise options like VPC and HIPAA, appealing to product and engineering teams that include non-technical contributors. Pricing is largely sales-led, and self-hosting is limited to Enterprise.

Vellum bundles the core steps of building an LLM feature into one platform: drafting and comparing prompts, building agents and workflows on a visual canvas with nodes for model calls, code, branching, retrieval, and human review, then evaluating outputs against datasets and deploying to production with monitoring and version control. In early 2026 the company repositioned toward 'Vellum for Agents,' emphasizing conversational agent building. It supports many model providers, avoiding lock-in, and offers enterprise-grade options like VPC deployment, HIPAA compliance, and self-hosting on its top tier. Vellum is a commercial SaaS, not open source. Its pricing is largely sales-led, with a limited free plan and higher tiers that route to a contact form, so specific dollar figures for mid-tier plans are not reliably public. It is used by product and engineering teams, including in regulated sectors like healthcare, that want a collaborative build-evaluate-deploy workflow.

TL;DR

Vellum is a commercial platform for building, evaluating, and deploying LLM apps and AI agents, combining prompt engineering, a visual workflow builder, evaluation suites, and production monitoring. It supports many providers and offers enterprise options like VPC and HIPAA, appealing to cross-functional teams. Pricing is largely sales-led with a limited free tier, and self-hosting is Enterprise-only. It raised a $20M Series A in 2025.

Company overview

Vellum is a New York-based company building a development platform for LLM applications. In early 2026 it repositioned toward agents, emphasizing conversational agent building for product and engineering teams.

It serves tech startups and regulated sectors such as healthcare and education, positioning itself as an all-in-one platform where technical and non-technical contributors can collaborate.

Product features

Vellum offers prompt management and a playground, a visual workflow and agent builder with nodes for model calls, code, branching, retrieval, tools, and human review, evaluation suites with datasets and model comparison, and production deployment with monitoring, versioning, and no-code prompt updates.

It supports many model providers including OpenAI, Anthropic, Google, and Cohere, plus hosted options via Azure OpenAI, Fireworks, and others, and is available on AWS Marketplace with an SDK and API.

Target market

Product and engineering teams, including less-technical stakeholders, building LLM features and agents; tech startups plus regulated sectors like healthcare, education, and customer support.

Buyer personas

End users

AI product builders, prompt engineers, and developers, plus non-technical collaborators contributing to agent and workflow design.

Buyers

Engineering and product leaders selecting an all-in-one LLM development platform.

Key influencers

Compliance and security stakeholders in regulated sectors, and platform engineers.

Ideal customer profile

A cross-functional product team building LLM features or agents that wants prompt, evaluation, and deployment in one platform, with enterprise deployment options when needed.

Funding & performance

Vellum has raised roughly $25.5M in total. Its most recent round was a $20M Series A announced in July 2025, led by Leaders Fund, with participation from Y Combinator, Socii Capital, Rebel Fund, Pioneer Fund, and Eastlink Capital.

Pros & cons

Pros

Combines prompt, workflow, evaluation, and deployment in one platform
Visual workflow builder lowers the barrier for cross-functional collaborators
Solid evaluation and testing tooling for pre-production quality gates
Multi-provider model support avoids lock-in
Free tier lets teams trial without a credit card
Enterprise options like VPC, HIPAA, and self-hosting for regulated buyers
SDK and API for code-first workflows and CI integration

Cons

Pricing is opaque and sales-led, making budgeting hard
Not open source, and self-hosting is locked to Enterprise
Free tier is quite limited in credits, documents, and seats
Smaller company and funding than infrastructure incumbents
Frequent repositioning can mean shifting product focus
Can become a proprietary orchestration layer you are tied to
Public feature and pricing details are inconsistent across sources

Pricing plans

Free

~50 builder credits/mo
1 seat
Hosted agent apps
Knowledge base up to 20 documents
No credit card required

Team

Contact sales

Higher usage limits
More integrations
Team collaboration
Sales-led pricing

Enterprise

Custom

SSO and advanced security
VPC or self-hosted deployment
HIPAA compliance
Dedicated support

Key features

API

Team collaboration

Self-hosted

Integrations

OpenAI, Anthropic, Google, Cohere, Azure OpenAI, Fireworks, AWS Marketplace

Input types

text, document

Output types

text

Best For

collaborative agent building, prompt evaluation and testing, deploying LLM features, cross-functional AI teams

Compare key features

View all alternatives →

Feature

Vellum

Dify

Braintrust

Pricing

Freemium

Free plan

Yes

Free trial

API

Yes

Self-hosted

Yes

Team support

Yes

Frequently asked questions

Is Vellum open source?+

No. Vellum is a commercial SaaS. It publishes an SDK and API and offers self-hosted or VPC deployment, but only on the Enterprise tier.

How much does Vellum cost?+

Vellum has a free plan, but paid pricing is largely sales-led and not fully published. Mid-tier and Enterprise plans route to a contact form, so budget by contacting sales.

What can I build with Vellum?+

You can build conversational agents and multi-step workflows using a visual canvas with nodes for model calls, code, branching, retrieval, and human review, then evaluate and deploy them.

Which model providers does Vellum support?+

It supports providers including OpenAI, Anthropic, Google, and Cohere, plus hosted options via Azure OpenAI, Fireworks, and others, avoiding single-vendor lock-in.

Is Vellum suitable for regulated industries?+

Yes, on Enterprise. It offers VPC or self-hosted deployment and HIPAA compliance, which suits healthcare and other regulated buyers.