Arcade.dev

Name: Arcade.dev
Brand: Arcade.dev

AI Native

Ship agents. Not auth infrastructure.

Est 2023|Raised $72M|18 Ref Customers

LinkedIn 6K|Web traffic 38K|FTEs 34|HQ Burlingame, CA

MarketsAI Agent Tool Integration Platforms MCP Gateway AI Agent Authentication and Authorization

Published Oct 15, 2026

01Common complaints

Teams that are primarily prototyping
- For demonstration or prototype use cases where enterprise authorization and governance are not requirements, direct API integrations or prototyping-focused platforms may be faster and lower cost.
Teams that primarily need deterministic workflow automation (not an agent runtime)
- Workflow automation platforms such as n8n or Pipedream are often better suited for deterministic integration workflows and long-running operational logic without agent abstractions.
Organizations extremely sensitive to per-call fees and not requiring enterprise controls
- Competitors offering large included call volumes and lower marginal per-call pricing may be more cost-effective for straightforward, high-volume workloads.
Organizations that require publicly verifiable compliance documentation prior to procurement
- Public documentation for certifications beyond the vendor's SOC 2 claim was not found; procurement-heavy organizations should require certification evidence during evaluation.
Organizations preferring minimal vendor layers over a runtime
- Teams wishing to self-host pure MCP servers and implement only specific auth, logging, or policy features in-house may prefer to avoid an additional runtime abstraction.
Current limitations acknowledged or implied
- Some OAuth providers may not be available as default, requiring custom auth provider configuration for certain integrations.
- The MCP client ecosystem is evolving; client availability and official support timing vary and should be validated for specific client requirements.
Roadmap items
- No public roadmap with firm dates was found; buyers typically request a product roadmap during sales evaluation or a demo to confirm planned capabilities.
Scale limitations
- Public-facing rate limits and usage guidance are referenced in pricing and documentation (for example, documented per-minute rate limits); buyers should confirm limits and escalation pathways for production scale.
Security/compliance gaps
- Beyond a vendor SOC 2 claim and enterprise features, a full compliance mapping and public attestations for ISO or HIPAA were not found in public materials.
API limitations or integration constraints
- Independent user reports of consistent API limitations were limited in the accessible sources; the risk profile is typical for emerging infrastructure platforms and should be assessed in a proof-of-concept.

- Teams that are primarily prototyping - For demonstration or prototype use cases where enterprise authorization and governance are not requirements, direct API integrations or prototyping-focused platforms may be faster and lower cost. - Teams that primarily need deterministic workflow automation (not an agent runtime) - Workflow automation platforms such as n8n or Pipedream are often better suited for deterministic integration workflows and long-running operational logic without agent abstractions. - Organizations extremely sensitive to per-call fees and not requiring enterprise controls - Competitors offering large included call volumes and lower marginal per-call pricing may be more cost-effective for straightforward, high-volume workloads. - Organizations that require publicly verifiable compliance documentation prior to procurement - Public documentation for certifications beyond the vendor's SOC 2 claim was not found; procurement-heavy organizations should require certification evidence during evaluation. - Organizations preferring minimal vendor layers over a runtime - Teams wishing to self-host pure MCP servers and implement only specific auth, logging, or policy features in-house may prefer to avoid an additional runtime abstraction. - Current limitations acknowledged or implied - Some OAuth providers may not be available as default, requiring custom auth provider configuration for certain integrations. - The MCP client ecosystem is evolving; client availability and official support timing vary and should be validated for specific client requirements. - Roadmap items - No public roadmap with firm dates was found; buyers typically request a product roadmap during sales evaluation or a demo to confirm planned capabilities. - Scale limitations - Public-facing rate limits and usage guidance are referenced in pricing and documentation (for example, documented per-minute rate limits); buyers should confirm limits and escalation pathways for production scale. - Security/compliance gaps - Beyond a vendor SOC 2 claim and enterprise features, a full compliance mapping and public attestations for ISO or HIPAA were not found in public materials. - API limitations or integration constraints - Independent user reports of consistent API limitations were limited in the accessible sources; the risk profile is typical for emerging infrastructure platforms and should be assessed in a proof-of-concept.

02Why they buy

unique selling points supported by available evidence
- A runtime framing that integrates authorization, tool execution reliability, and governance in a single managed layer.
- A usage-based pricing model including a free tier for initial experimentation.
- Deployment flexibility with vendor-managed cloud options and self-hosted Kubernetes deployment pathways described in documentation.
- Open-source components (for example, an arcade-mcp GitHub project) that enable custom MCP server implementations.
unique selling points requiring further validation
- Claims about very large numbers of available tools (vendor-published counts vary across pages and directories).
- Reliability features such as parallelized execution and intelligent retries are described in marketing materials and require independent validation in production settings.
Areas where competitors may have advantages
- Competitors may offer broader integration catalogs and lower effective per-call pricing for prototyping and high-volume simple workloads.
- Workflow automation tools may provide richer tooling for deterministic process automation.

- unique selling points supported by available evidence - A runtime framing that integrates authorization, tool execution reliability, and governance in a single managed layer. - A usage-based pricing model including a free tier for initial experimentation. - Deployment flexibility with vendor-managed cloud options and self-hosted Kubernetes deployment pathways described in documentation. - Open-source components (for example, an arcade-mcp GitHub project) that enable custom MCP server implementations. - unique selling points requiring further validation - Claims about very large numbers of available tools (vendor-published counts vary across pages and directories). - Reliability features such as parallelized execution and intelligent retries are described in marketing materials and require independent validation in production settings. - Areas where competitors may have advantages - Competitors may offer broader integration catalogs and lower effective per-call pricing for prototyping and high-volume simple workloads. - Workflow automation tools may provide richer tooling for deterministic process automation.

Why customers chose Arcade (based on vendor and community signals)

Buyers seeking multi-user, per-user OAuth/permissions without building a custom auth infrastructure.
Buyers needing centralized governance and auditability that attribute actions to users and agents.
Buyers interested in tool implementations designed for agent workflows rather than plain OpenAPI wrappers, per vendor and community signals.

03Field reports

Important data-quality note:

The name "Arcade" is used by several unrelated products; multiple unrelated projects may be present in third-party directories and review aggregators. Some review listings may conflate or mis-categorize the Arcade.dev MCP runtime with other products, which introduces noise.

Direct quotes (vendor testimonials on the vendor site)

Vendor-published testimonial (LangChain CEO): Vendor-provided endorsement states that the runtime makes the MCP offering enterprise-ready; this is a vendor-curated endorsement rather than an independent review.
Additional vendor-published endorsements on the homepage emphasize secure MCP, integration behavior, and token handling; these are vendor-curated.

Review platforms (attempted and findings)

G2: An Arcade.dev listing exists with competitor/alternatives pages and a pricing overview; some pages indicate inconsistent or incomplete pricing disclosures relative to the vendor's pricing page. Star ratings cited on third-party directories should be validated by reading the underlying reviews directly.
Gartner Peer Insights: No clearly accessible Gartner Peer Insights product page for the Arcade.dev MCP runtime was retrieved in accessible results; the platform often requires login or blocks automated access.
TrustRadius: No distinct TrustRadius review page for this specific MCP runtime was found in accessible results.

- G2: An Arcade.dev listing exists with competitor/alternatives pages and a pricing overview; some pages indicate inconsistent or incomplete pricing disclosures relative to the vendor's pricing page. Star ratings cited on third-party directories should be validated by reading the underlying reviews directly. - Gartner Peer Insights: No clearly accessible Gartner Peer Insights product page for the Arcade.dev MCP runtime was retrieved in accessible results; the platform often requires login or blocks automated access. - TrustRadius: No distinct TrustRadius review page for this specific MCP runtime was found in accessible results.

Community/Social (Reddit) quotes and themes

Market positioning confusion: Community threads compare Arcade.dev to other MCP and agent-runtime projects (e.g., Composio), indicating that differentiation in the market is still an open topic.
Pricing discussion: Community threads contrast per-call pricing across vendors, and note that comparisons often omit factors such as token efficiency, tool behavior, and retry rates.
Product announcement activity: Some community posts appear to be authored by individuals affiliated with the vendor; these should be discounted as independent validation.

- Market positioning confusion: Community threads compare Arcade.dev to other MCP and agent-runtime projects (e.g., Composio), indicating that differentiation in the market is still an open topic. - Pricing discussion: Community threads contrast per-call pricing across vendors, and note that comparisons often omit factors such as token efficiency, tool behavior, and retry rates. - Product announcement activity: Some community posts appear to be authored by individuals affiliated with the vendor; these should be discounted as independent validation.

Why some customers might leave / churn (inferred from critiques)

Teams that are strictly prototyping or hobbyist and do not require enterprise governance may find the platform adds overhead.
Very high-volume, simple workloads may favor alternatives with lower per-call pricing.
Procurement processes may be delayed if specific compliance attestations or detailed security documentation are required but not publicly available.

Case Studies

Vendor-published benchmark: Arcade vs Composio token efficiency (Attio toolkit)
- Vendor-published benchmark (recent) compared an Attio MCP toolkit implementation to a Composio implementation across CRM queries and reported a large token-efficiency delta in that specific test case.
- Critical evaluation:
  - This is a vendor-published benchmark; while reproducibility artifacts are provided in vendor materials, the selection of queries, toolkit implementations, and representativeness across other workloads may introduce bias.
  - Even if token efficiency is observed in the benchmark, buyer value depends on workload similarity, alternative configuration options for competitors, and operational behaviors such as retry rates.
Vendor whitepapers/pdfs
- The vendor publishes whitepapers arguing that MCP benefits from a runtime architecture and comparing runtime approaches to gateways.
- Critical evaluation:
  - These documents are vendor-positioning materials and provide perspective on the vendor's architectural choices; they are not independent analyst validations.

- Vendor-published benchmark: Arcade vs Composio token efficiency (Attio toolkit) - Vendor-published benchmark (recent) compared an Attio MCP toolkit implementation to a Composio implementation across CRM queries and reported a large token-efficiency delta in that specific test case. - Critical evaluation: - This is a vendor-published benchmark; while reproducibility artifacts are provided in vendor materials, the selection of queries, toolkit implementations, and representativeness across other workloads may introduce bias. - Even if token efficiency is observed in the benchmark, buyer value depends on workload similarity, alternative configuration options for competitors, and operational behaviors such as retry rates. - Vendor whitepapers/pdfs - The vendor publishes whitepapers arguing that MCP benefits from a runtime architecture and comparing runtime approaches to gateways. - Critical evaluation: - These documents are vendor-positioning materials and provide perspective on the vendor's architectural choices; they are not independent analyst validations.

04Best fit

Company sizes
- Best fit appears to be teams transitioning from prototype to production agent deployments and enterprises requiring identity integration and auditability.
- Typical organization sizes: mid-sized to large organizations where multi-user AI agents are being deployed across systems.
Industries and verticals
- Regulated or security-sensitive industries (finance, healthcare, enterprise SaaS) where agent actions must be attributable and controlled.
- Organizations with significant SaaS footprints (Google Workspace, Slack, Salesforce, GitHub, Jira/Confluence, etc.).
Roles and personas
- Decision makers:
  - Head of Platform Engineering, CTO, VP Engineering
  - Security/IT leadership (CISO organization)
- Practitioners:
  - AI/ML engineers and agent developers
  - Internal tools and platform engineering teams
Pain points the product addresses
- Multi-user OAuth and credential management separate from model inputs.
- Tool reliability concerns (schema ambiguity, parameter hallucinations, and noisy responses) through structured tool definitions.
- Governance needs such as audit logs and policy controls for agent actions.
Qualification criteria (good fit)
- Building agent systems that must take real actions across business systems.
- Requiring user-scoped authorization tied to enterprise identity.
- Requiring audit trails and approval processes for security review.
Deal disqualifiers (poor fit)
- Single-user scripts, hobby projects, or offline-only agents where runtime governance is unnecessary.
- Organizations that already maintain a robust internal integration and auth platform and require only minimal MCP plumbing.
- Organizations that require specific compliance attestations (for example, ISO 27001 or HIPAA) and cannot proceed without rapid access to formal documentation.

- Company sizes - Best fit appears to be teams transitioning from prototype to production agent deployments and enterprises requiring identity integration and auditability. - Typical organization sizes: mid-sized to large organizations where multi-user AI agents are being deployed across systems. - Industries and verticals - Regulated or security-sensitive industries (finance, healthcare, enterprise SaaS) where agent actions must be attributable and controlled. - Organizations with significant SaaS footprints (Google Workspace, Slack, Salesforce, GitHub, Jira/Confluence, etc.). - Roles and personas - Decision makers: - Head of Platform Engineering, CTO, VP Engineering - Security/IT leadership (CISO organization) - Practitioners: - AI/ML engineers and agent developers - Internal tools and platform engineering teams - Pain points the product addresses - Multi-user OAuth and credential management separate from model inputs. - Tool reliability concerns (schema ambiguity, parameter hallucinations, and noisy responses) through structured tool definitions. - Governance needs such as audit logs and policy controls for agent actions. - Qualification criteria (good fit) - Building agent systems that must take real actions across business systems. - Requiring user-scoped authorization tied to enterprise identity. - Requiring audit trails and approval processes for security review. - Deal disqualifiers (poor fit) - Single-user scripts, hobby projects, or offline-only agents where runtime governance is unnecessary. - Organizations that already maintain a robust internal integration and auth platform and require only minimal MCP plumbing. - Organizations that require specific compliance attestations (for example, ISO 27001 or HIPAA) and cannot proceed without rapid access to formal documentation.

05The pitch

What the product does (core value proposition in objective terms)

Arcade positions itself as an MCP runtime (Model Context Protocol runtime) for production AI agents: an intermediary layer that provides (a) agent/user authorization, (b) a catalog of agent-oriented tools/integrations, and (c) governance and auditability for tool execution.
Stated rationale: some agent projects encounter production challenges around authorization, reliability of tool-calling, and governance; Arcade offers these capabilities as a managed runtime layer to address those gaps.

- Arcade positions itself as an MCP runtime (Model Context Protocol runtime) for production AI agents: an intermediary layer that provides (a) agent/user authorization, (b) a catalog of agent-oriented tools/integrations, and (c) governance and auditability for tool execution. - Stated rationale: some agent projects encounter production challenges around authorization, reliability of tool-calling, and governance; Arcade offers these capabilities as a managed runtime layer to address those gaps.

Target market and market positioning

Target buyers:
- Engineering/platform teams deploying multi-user AI agents into production environments.
- Security and compliance stakeholders responsible for approving agent actions performed on behalf of users.
Positioning:
- Presented as an agent tool execution layer intended for production deployments rather than prototype tool wrappers.
- Emphasis on identity provider (IdP) integration and least-privilege / per-user authorization.

- Target buyers: - Engineering/platform teams deploying multi-user AI agents into production environments. - Security and compliance stakeholders responsible for approving agent actions performed on behalf of users. - Positioning: - Presented as an agent tool execution layer intended for production deployments rather than prototype tool wrappers. - Emphasis on identity provider (IdP) integration and least-privilege / per-user authorization.

Get a comprehensive analysis of Arcade.dev including market position, competitive landscape, adoption trends, and peer benchmarks.

Download full report →