Does Inngest support MCP and tool calling?

Full under the Agentic Index verification standard. Inngest publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: First party SDKs across TypeScript, Python and Go with an open source core, AgentKit as an extensible framework, and a local development server, so the entire product surface is code addressable. No MCP server is documented. 2026-08-05. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Inngest have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. Inngest publishes public evidence of enterprise security and identity controls, covering a compliance posture such as SOC 2 or ISO 27001, single sign on (SSO or SAML), and role based access control (RBAC). Evidence basis: SOC 2, a HIPAA business associate agreement, end to end encryption and single sign on with SAML, with dedicated infrastructure available on enterprise tiers. 2026-08-05. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can Inngest be self hosted or run in a private VPC?

Full under the Agentic Index verification standard. Inngest publishes public evidence of self hosted or private VPC deployment, along with data residency options for controlling where agent data is processed and stored. Evidence basis: Architecturally distinctive: Inngest orchestrates but does not host, calling developer owned HTTP endpoints on serverless or traditional hosts, so compute stays on customer infrastructure across Vercel, Cloudflare Workers, AWS Lambda or any. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Inngest

Q: Who is Inngest for?

backend and full stack developers, AI engineering teams, growth stage SaaS and AI startups

Q: What are the best alternatives to Inngest?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Inngest among agent infrastructure platforms are Trigger.dev, DBOS, Windmill, Bernstein. Each is graded on the same public evidence, and no vendor pays for placement.

Also known as: AgentKit, step.ai, Inngest AgentKit

Visit site

Agent infrastructureindependentVerified 2026-08-05

Entry priceFree tier reported at 50,000 runs per month; Basic around $30 per month; Pro around $150 per month; enterprise customFull pricing detail

Durable execution platform for event driven workflows and AI agents, with step level retries, checkpointed resume, native human in the loop waits, and an AgentKit framework for multi agent networks, running on your own compute.

Inngest is a durable execution platform for event driven workflows, background jobs and AI agents. Developers write ordinary TypeScript, Python or Go functions, wrap external calls in step primitives, and the orchestration engine handles queuing, state persistence, retries with exponential backoff, checkpointing so a failure resumes where it stopped, scheduling, concurrency limits, parallel fan out and fan in, batching, debounce, throttling and prioritisation. None of that requires additional infrastructure.

For agents specifically it ships AgentKit, a framework for composing single agents or networks of agents that collaborate, with tool use, routing, stateful orchestration where network state stays available to routers, agents and tools, and lifecycle hooks for adjusting prompts and inspecting outputs mid execution. Because it sits on the same durable substrate, an agent crash partway through a loop does not drop the user's request. A companion feature, step.ai, wraps model calls in the same reliability guarantees.

The capability worth singling out is step.waitForEvent, which pauses a workflow for hours or days waiting on an external event. That makes human in the loop a native execution primitive rather than an application concern, which is a materially different position from platforms where waiting on a human means building your own state machine.

Operationally it is unusual in that Inngest orchestrates but does not host: in production it calls developer owned HTTP endpoints on serverless or traditional hosts, so compute stays on the customer's infrastructure across Vercel, Cloudflare Workers, AWS Lambda or any container. A one command local development server gives production parity with hot reloading and a visual trace UI, and production carries step level tracing, logs and bulk Replay for recovering from failures after the fact. Security covers SOC 2, a HIPAA business associate agreement, end to end encryption and single sign on with SAML. The category's incumbent is Temporal, which raised $300 million at a $5 billion valuation in February 2026; Inngest's position is durable agent workflows with no stateful infrastructure to operate.

Vendor details

Canonical URL

https://www.inngest.com

Company status

independent

Use cases & customers

Target customers

backend and full stack developersAI engineering teamsgrowth stage SaaS and AI startups

Deployment options

SaaSserverlesscustomer-owned computededicated infrastructure

In practice

Your agent loop runs for twenty minutes across a dozen model and tool calls and a single rate limit kills the whole request. Step level checkpointing resumes from the failure point rather than restarting the run.

A workflow needs to stop and wait for a human to approve something, possibly for days. step.waitForEvent makes that a native execution primitive instead of a state machine you build and maintain yourself.

You cannot move compute off your own infrastructure. Inngest orchestrates by calling your own HTTP endpoints on Vercel, Lambda, Cloudflare Workers or a container, so execution stays where your code already lives.

Sources & related URLs

Research sources

https://www.inngest.com/ai https://www.inngest.com/blog/ai-orchestration-with-agentkit-step-ai https://devtune.ai/verticals/open-source-commercial-oss-infrastructure/inngest https://aitoolsatlas.ai/tools/inngest/review

Agentic Index coverage score

10.5 / 14 capabilities · 75%

Integrations & Tool CallingAgentKit makes tool use a first class primitive so agents can act rather than only generate text, and functions connect to any API, service or database through code, but there is no prebuilt connector catalogue or managed authentication layer. 2026-08-05	Partial
Workflow OrchestrationThe core product: step based durable orchestration with automatic retries and exponential backoff, checkpointed resume, concurrency limits, parallel fan out and fan in, batching, debounce, throttling and prioritisation, plus AgentKit for composing networks of collaborating agents with routing. 2026-08-05	Full
Knowledge Grounding & RAGNo retrieval, vector store or knowledge grounding capability is provided. RAG pipelines are a documented workload that runs ON Inngest, but the grounding layer itself belongs to the application. 2026-08-05	Unable to verify
Human Oversight & Guardrailsstep.waitForEvent pauses a workflow for hours or days awaiting an external signal, making human in the loop a native execution primitive rather than an application concern, and lifecycle hooks allow inspecting and adjusting agent behaviour mid run. 2026-08-05	Full
Security, Identity & GovernanceSOC 2, a HIPAA business associate agreement, end to end encryption and single sign on with SAML, with dedicated infrastructure available on enterprise tiers. 2026-08-05	Full
Observability & AuditabilityStep level tracing, execution timeline visualisation, logs and a dashboard in production, plus bulk Replay to recover and re run failed executions after the fact. Observability is a first class surface rather than an add on. 2026-08-05	Full
Memory & State PersistenceState persistence is the product. The engine checkpoints every step so execution resumes after crashes or restarts, and AgentKit keeps network state available to routers, agents and tools across a multi agent run. 2026-08-05	Full
Deployment & Data ResidencyArchitecturally distinctive: Inngest orchestrates but does not host, calling developer owned HTTP endpoints on serverless or traditional hosts, so compute stays on customer infrastructure across Vercel, Cloudflare Workers, AWS Lambda or any container, with dedicated infrastructure available on enterprise. 2026-08-05	Full
Prebuilt Agents, Templates & PacksAgentKit supplies agent and network composition primitives and there are documented patterns, but it is a framework rather than a catalogue: no prebuilt agents, template gallery or ready to deploy packs are documented. 2026-08-05	Partial
Triggers & Channel CoverageEvent driven invocation is the architecture, with real time event triggers, built in scheduling that removes external cron, event batching and debouncing, and step.waitForEvent for resuming on an external signal. 2026-08-05	Full
Model Flexibility & Routingstep.ai wraps model calls generically and the platform is model agnostic by construction, orchestrating whichever provider the developer calls, but no model list, routing policy, fallback chain or selection interface is offered by the platform itself. 2026-08-05	Partial
APIs, SDKs & MCP ExtensibilityFirst party SDKs across TypeScript, Python and Go with an open source core, AgentKit as an extensible framework, and a local development server, so the entire product surface is code addressable. No MCP server is documented. 2026-08-05	Full
Testing, Debugging & OptimizationA one command local development server with production parity and hot reloading, a visual trace and execution timeline for debugging, step level inspection, and bulk Replay to re run failed executions. Substantively more debugging capability than the norm in this index, though no agent quality evaluation harness or benchmark suite is documented. 2026-08-05	Full
Browser & Computer UseNo browser control or computer use capability. The platform executes developer written functions and API calls rather than operating interfaces. 2026-08-05	Unable to verify

The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Recent platform changes

2026-08-10·Workflow orchestrationVerified

Inngest released a feature allowing developers to rerun workflows from any specific step, even if the run has successfully completed. The platform reconstructs all preceding steps as memoized state and only re-executes the selected step and subsequent operations.

Bears on: Memory / state

View source

View all 1 change for Inngest →Tracked since Aug 2026 · Verified from public vendor sources

Pricing

Free tier reported at 50,000 runs per month; Basic around $30 per month; Pro around $150 per month; enterprise custom

runs and steps executed

Free tierTrial available

What is public

Tier structure and approximate rates circulate widely in third party reviews and Inngest publishes a pricing page, though it was not retrieved in this pass.

Billing mechanics

Published self serve tiers metered on execution volume, with a free tier reported at 50,000 runs a month, a Basic tier near $30 and a Pro tier near $150. Enterprise tiers add SOC 2 assurances and dedicated infrastructure. Because Inngest calls customer owned endpoints, the compute cost of running the work itself sits on the customer's own hosting bill rather than in the subscription.

Cost watchouts

Billing counts steps, and an agent loop generates far more steps per run than a background job, so a run based mental model materially understates cost for agent workloads

Variable cost rationale

Execution metering ties cost to workload shape, and agent loops produce many more steps per run than background jobs, which is the documented way spend outruns expectations. Offsetting that, compute runs on customer infrastructure so the subscription covers orchestration rather than execution.

Additional watchouts

Independent comparisons single out step counts as the pricing trap: agent workloads fan out into many steps per run, so the effective bill can diverge sharply from a run based estimate. Total cost also splits across two bills, the Inngest subscription and the customer's own compute hosting.

Overage / add-ons

Consumption is metered on executions; higher volume moves the customer up a tier, with enterprise offering dedicated infrastructure

Sales call required

Mixed (some tiers require a call)

Free / trial

Free tier reported at 50,000 runs per month, generous enough to run real production workloads before paying

Lowest paid plan

Basic

Commercial notes

Published pricing is a deliberate contrast with Temporal, the category incumbent, whose cloud pricing and self hosted operational burden are both harder to model. Independent comparison of the two frames it as Temporal for polyglot mission critical durability at scale versus Inngest for shipping durable TypeScript agents quickly with nothing stateful to operate.

Key ambiguities

All figures come from third party reviews rather than a retrieved first party pricing page, and are described as approximate. The exact relationship between runs, steps and plan limits is not confirmed. Re verify on inngest.com before publishing any number.

Missing data

Confirmed first party rates, step allowances per tier, the run to step accounting relationship, and enterprise pricing.

Agentic Index verified 2026-08-05

Official pricing page

Related vendors

Alternatives to Inngest

The closest documented capability profiles to Inngest among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

Trigger.dev10.5 / 14Adds documented Browser & Computer Use
DBOS9.0 / 14A lighter documented profile than Inngest
Windmill10.0 / 14Fuller documented coverage on Integrations & Tool Calling
Bernstein10.5 / 14Fuller documented coverage on Integrations & Tool Calling and Model Flexibility & Routing
Temporal9.5 / 14Fuller documented coverage on Integrations & Tool Calling
Tray.ai11.5 / 14Adds documented Knowledge Grounding & RAG

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Compare Inngest with other vendors →Compare all agent infrastructure platforms →See where Inngest ranks →

Data confidence: high