AWS Lambda runs your functions. Avalanche runs the work your agents do.

Quick answer

Updated October 27, 2026. AWS Lambda is general serverless compute: a single invocation runs for up to 15 minutes, and multi-step workflows are built with Lambda durable functions, which checkpoint and replay for up to a year, or with AWS Step Functions.[1, 5, 16] Both AWS and Avalanche let you return to a previous version without a rebuild, and both keep running work on the version it started on.[2, 7, 14, 25] Choose AWS Lambda if developers are building application backends on AWS and need its scale, IAM, CloudTrail, budgets and 220+ service integrations.[18, 34, 35, 36] Choose Avalanche if the job is a long, multi-step, document-heavy business workflow: you set the few steps, checks and approvals that matter, agents do the work inside each step, and every run keeps typed step outputs and versioned files your team can compare across runs and versions.[21, 23, 25]

Key takeaway

Choose Avalanche when agents should do the work inside a few business steps you control, often over PDFs, Word files and spreadsheets.

Choose AWS Lambda when developers are building application backends on AWS and need its scale, identity controls and service integrations.

  • Avalanche wins on level of abstraction: you set the steps, checks and approvals; an agent does the work inside each step instead of functions and state machines you assemble yourself.
  • Avalanche wins on rolling back a whole workflow at once: AWS rolls back one function or state machine at a time by moving its alias; Avalanche Cloud versions the whole workflow as one deploy and puts every step back in one action, with no rebuild, while running work stays on its version.
  • AWS Lambda wins on scale and integrations: Lambda durable functions allow millions of running executions per Region, and Step Functions integrates with 220+ AWS services; Avalanche Cloud is on a waitlist.
  • AWS Lambda wins on identity and access: IAM policies down to a function version or alias, CloudTrail records of every Lambda API call, and AWS Budgets with alerts and actions; Avalanche has no spend limits yet.

By , CTO, Trampoline AI. Updated · Facts checked · Published by the Avalanche team. Avalanche competes with AWS Lambda.

01 / The core differenceAvalanche vs AWS Lambda

Serverless compute you assemble, or agentic workflows for your business.

AWS Lambda runs the functions developers write, and AWS documents two ways to chain them into long workflows: durable functions in code and Step Functions state machines.[1, 18] Avalanche works a level higher: you describe the outcome and the few steps, checks and approvals that matter, and agents do the work in between, with each step checkpointed.[21, 25]

AWS Lambda is general-purpose compute for developers: you write functions, and for multi-step work you chain them with durable functions in code or a Step Functions state machine, then wire in schedules, storage, permissions and monitoring from other AWS services.[1, 18, 30, 33] Avalanche is built for the work agents do for a business: you define the few steps, checks and approvals that matter, agents do the work in between, and every run keeps its version, its typed step outputs and the files each step changed.[21, 23, 25] Both record every execution; AWS execution history and metrics help a developer debug and operate code, while Avalanche step outputs let a team compare the results of agent work across many runs and versions.[19, 25]

02 / At a glanceChecked 2026-10-27

Avalanche and AWS Lambda, side by side.

Avalanche and AWS Lambda compared, checked 2026-10-27
CriterionAvalancheAWS LambdaEdge
Level of abstractionCoarse business steps: you set the steps, checks and approvals, and an agent does the work inside each step.[20, 25]A developer writes each function, then chains them with durable-function code or a Step Functions state machine in Amazon States Language.[1, 18]Avalanche edge
Agents and skills inside stepsAgent steps can run Claude through the Claude Agent SDK and load Agent Skills (SKILL.md); the Avalanche skill drafts a full workflow from a described outcome.[22, 26, 27]Durable functions are pitched for AI workflows that chain model calls; the agent loop, tools and model calls are code you write inside steps. A skill loader inside steps is not documented.[1]Avalanche edge
Document lineageAvalanche Cloud: every run gets a versioned workspace; each step's file changes are committed as a new version, so you can see which step changed a file, compare versions and download any step's files; agent steps get PDF, Word and spreadsheet skills.[23, 25]Temporary /tmp storage per execution environment (512 MB to 10,240 MB), EFS or S3 for files; durable checkpoints hold up to 1 MB per operation, and AWS advises external storage for large data. Per-step file versions are code you bring.[4, 5, 32, 33]Avalanche edge
Inspecting and comparing runsAvalanche Cloud: typed inputs and outputs at every step checkpoint, in one run history filtered by status, flow and deployment; a web console with a live DAG of every flow and run, the inputs and outputs of any step in one click, and the full console in your terminal, so you can compare a step across many runs and versions.[21, 25]Durable execution history kept 1 to 90 days, CloudWatch metrics, EventBridge status events and X-Ray traces; Step Functions shows each execution visually and keeps history 90 days after it closes.[4, 16, 19, 40]Different jobs
Long runs and recoveryAvalanche Cloud: workflows run for hours or days; completed steps are recorded durably and never re-run after a crash; failed steps retry, and uncertain outcomes are held instead of blindly retried.[21, 25]One invocation runs up to 15 minutes. Durable functions checkpoint each step and replay for up to a year, so step code must replay safely; Step Functions Standard runs up to a year and redrives from the failed step.[1, 3, 5, 16, 17, 29]Similar
Versions and rollbackAvalanche Cloud: each deploy is a frozen version of the whole workflow, mapped to an exact git commit, with locked, hash-verified dependencies; running workflows finish on the version they started on; instant rollback to any previous version from the CLI or console, no rebuild; an audit trail of who promoted or rolled back.[25]Each Lambda function and each state machine has its own immutable versions and aliases; repointing an alias rolls back with no rebuild, and durable and Step Functions executions keep the version they started on, unless started on $LATEST or an unqualified ARN. A workflow built from several functions and a state machine is rolled back resource by resource.[2, 6, 7, 11, 12, 13, 14]Avalanche edge
Scale, integrations and availabilityAvalanche Cloud: FIFO queueing when over capacity, with organization and worker-pool limits; Avalanche Cloud is on a waitlist, and the self-hostable MIT-licensed runtime is more limited than Cloud.[20, 24, 25, 28]Generally available; millions of running durable executions per Region; Step Functions integrates with 220+ AWS services and 16,000 APIs.[5, 16, 18]AWS Lambda edge
Schedules, webhooks, CLI and APIAvalanche Cloud: cron schedules and webhooks declared in your workflow code, following rollbacks and promotions; webhook URLs never change on redeploy, redeliveries are deduplicated for 7 days, and idempotency keys return the same run for 7 days; a typed REST API with an OpenAPI spec, and a CLI.[25]EventBridge Scheduler cron, rate and one-time schedules; function URLs whose endpoint never changes, usable as webhooks; event sources such as SQS, Kinesis and DynamoDB; idempotent durable starts by execution name; the AWS CLI, SDKs and APIs.[2, 30, 31]Similar
Identity and accessAvalanche Cloud: SSO; organization roles and per-action permissions, managed in your organization's settings; scoped, encrypted secrets; agents see only what each step passes them (its inputs, its files and the secrets it declares); steps run non-root in a sandbox, cut off from cloud metadata and private networks; an audit trail of promotions and rollbacks; no spend limits yet.[25]IAM identity and resource policies down to a function version or alias; CloudTrail records every Lambda API call with the identity behind it; AWS Budgets alerts on spend and can apply a deny policy; customer-managed KMS keys for durable data.[4, 6, 34, 35, 36, 41]AWS Lambda edge
03 / Where each one winsOne claim per card

Where each one wins

Where Avalanche wins

  • Higher-level workflows

    You set the few business steps, checks and approvals that matter; an agent does real work inside each step. Python steps keep exact rules and arithmetic.[20, 25]

  • From a skill to a workflow

    Install the Avalanche skill with npx skills add Trampoline-AI/avalanche, describe the outcome in your coding agent, and it drafts the workflow with you.[22]

  • Every step versions your documents

    Avalanche Cloud gives every run a versioned workspace. Pass one document or hundreds; each step works on the files directly, and every step's changes are committed as a new version of the workspace, so you can see which step produced or changed each file, compare versions, and download any step's files. Document lineage, the way data teams have table lineage. Agent steps get PDF, Word and spreadsheet skills.[23, 25]

  • One version for the whole workflow

    On Avalanche Cloud each deploy is a frozen, checksum-verified version of the whole workflow, mapped to an exact git commit. Running workflows finish on the version they started on, and instant rollback makes any previous version live again in one click or one command, with no rebuild. Avalanche Cloud is on a waitlist.[25, 28]

  • Runs your business team can compare

    Each step is a business result with typed outputs, such as an RFP answer or a contract finding. Avalanche Cloud keeps every run by workflow and version, so you can compare one step across many runs.[21, 25]

  • Checkpointed Claude agent steps

    Agent steps can run Claude through the Claude Agent SDK and load Agent Skills (SKILL.md), each sandboxed and checkpointed like any other step.[26, 27]

Where AWS Lambda wins

  • Durable functions for up to a year

    Lambda durable functions checkpoint each step and replay after an interruption, skipping completed work; an execution can run for up to one year, and waits suspend it without compute charges.[1, 3]

  • Scale

    Durable functions allow 5,000,000 running executions per Region by default (10,000,000 in the largest Regions), and Step Functions allows 1,000,000 open executions per account and Region.[5, 16]

  • The AWS ecosystem

    Step Functions orchestrates 220+ AWS services and 16,000 APIs, with a visual builder; Lambda takes events from SQS, Kinesis, DynamoDB, S3, SNS, EventBridge and more.[2, 18, 37]

  • Rollback and traffic shifting per function

    Repoint a Lambda alias to an earlier published version from the console, CLI or API, split traffic between two versions, or let CodeDeploy shift traffic in canary or linear steps and roll back automatically on an alarm.[7, 8, 9, 10]

  • Identity, audit and budgets

    IAM controls who can call or change each function, version and alias; CloudTrail records Lambda API calls with the identity behind them; AWS Budgets alerts on spend and can apply a deny policy.[6, 34, 35, 36]

  • Generally available and pay per use

    Lambda bills requests and duration, with a monthly free tier; durable functions add charges for durable operations and stored data, and waits incur no duration charge.[37]

04 / Detailed comparison4 buyer questions

What changes for your team.

01

Serverless compute or business workflows?

Buyer question

Is the work part of your application's backend on AWS, or a business process your team repeats every week?

AWS Lambda runs functions in response to events and scales them for you. A single invocation is capped at 15 minutes, so AWS documents two ways to build longer, multi-step work: durable functions, where you wrap calls in checkpointed steps in JavaScript, TypeScript, Python, Java, Go or C#, and Step Functions, a separate service with a graph-based language and visual builder.[1, 5, 18] AWS itself says durable functions suit application logic inside Lambda and Step Functions suits orchestration across AWS services.[18]

Avalanche is built for long, multi-step business work that agents do, such as drafting RFP answers from past proposals or reviewing a stack of contracts. You set the few steps, checks and approvals that matter, agents do the work in between, and simple workflows fit the same model. On Avalanche Cloud, workflows run for hours or days, and an owner approves results before they go out.[25]

02

Rollback: one function at a time, or the whole workflow at once

Buyer question

When a release goes wrong in production, how fast can you undo it, and what happens to running work?

On AWS, publishing a Lambda version freezes its code and most settings, and an alias is a pointer you can move to any published version from the console, CLI or API, with no rebuild; CodeDeploy can shift alias traffic in canary or linear steps and roll back automatically on an alarm.[6, 7, 9, 10] Step Functions works the same way for state machines: AWS documents rolling back by setting the alias to send 100 percent of executions to the earlier version.[11, 13] Running work stays put. A durable execution is pinned to the function version that started it, and after an alias update in-progress executions continue on their version; executions started on $LATEST are not pinned and can resume on new code.[2] A Step Functions execution is tied to its version when it starts, running executions keep the previous definition after an update, and a redrive keeps the original version even if the alias has moved.[14, 15, 17]

Avalanche Cloud does the same for the whole workflow at once. A deploy is one frozen, checksum-verified version of every step and its bundled dependencies; running workflows keep the version they started on, and you roll back to any previous version in one click or one command, with no rebuild, with an audit trail of who did it.[25] On AWS, each function and each state machine has its own version history, and a state machine task runs whichever function version, alias or $LATEST its ARN names, so a workflow built from several of them is rolled back resource by resource.[13, 41] Avalanche Cloud is on a waitlist.[28]

03

Documents, agents and skills

Buyer question

How much do you want to build around the agent yourself?

On Lambda, files live in temporary /tmp storage that is unique to each execution environment, on an EFS mount, or in S3; a durable checkpoint holds up to 1 MB per operation and 100 MB per execution, and AWS advises keeping state small and using external storage for large data.[4, 5, 32, 33] Agents are code you write: AWS pitches durable functions for AI workflows that chain model calls and pause for human feedback, and Step Functions has a documented human-approval pattern.[1, 39]

In Avalanche, files move between steps, and on Avalanche Cloud each step's file changes are committed as a new version of the run's workspace. Agent steps declare typed inputs and outputs and get PDF, Word and spreadsheet skills.[21, 23] The Avalanche skill drafts the workflow from an outcome you describe in your coding agent.[22] Agent steps can also run Claude through the Claude Agent SDK and load Agent Skills (SKILL.md).[26, 27]

04

Running it in production

Buyer question

What do you need from the platform around the workflow, and what do you need to see about each run?

AWS has the edge on scale, ecosystem and identity. Durable functions allow millions of running executions per Region, Step Functions reaches 220+ AWS services, and IAM, CloudTrail and AWS Budgets cover access, audit and spend.[5, 18, 34, 35, 36] Lambda bills requests and duration; durable functions add durable operations, data written and data retained; Step Functions Standard bills each state transition, retries included.[37, 38] Observability is built for developers: durable execution history, CloudWatch metrics per execution and per Step Functions version, EventBridge events and X-Ray traces.[14, 19]

You can self-host the MIT-licensed Avalanche runtime, with its operator, web console, REST API and terminal UI, on your own infrastructure; it is more limited than Avalanche Cloud, which runs the production infrastructure for you.[20, 24] Avalanche Cloud adds SSO and organization roles, versioned deploys, instant rollback of the whole workflow with no rebuild, step checkpoints, managed queueing and one run history for every deployed workflow, and is open through a waitlist.[25, 28] Both record each execution; the difference is what a step is. A Lambda step is code a developer wrote, so its history and traces help debug code. An Avalanche step is a business result an agent produced, with typed outputs and versioned files, so a team can compare that step across hundreds of runs and see which outputs drift or which version changed them.[21, 25]

05 / FAQDirect answers

Common questions.

Is Avalanche an alternative to AWS Lambda?
For agentic business workflows, yes; for general serverless compute, no. Lambda runs any function on events at large scale, and durable functions or Step Functions turn functions into long workflows.[1, 18] Avalanche runs business workflows where agents work inside a few steps you set, with versioned deploys and document lineage on Avalanche Cloud.[25]
How long can an AWS Lambda function run?
A single invocation runs for up to 900 seconds (15 minutes); functions on Lambda Managed Instances invoked asynchronously can run up to 90 minutes.[5] A durable function execution can span many invocations and run for up to one year, and Step Functions Standard workflows also run for up to one year.[3, 16]
What are Lambda durable functions?
Regular Lambda functions that use the AWS durable execution SDK. You wrap work in steps that checkpoint their results; after a pause or failure, Lambda replays your code from the beginning and skips completed steps, so step code must replay safely. Waits suspend the function without compute charges.[1, 3, 29]
Can I roll back a Lambda function or Step Functions workflow without redeploying?
Yes. Point a Lambda alias back at an earlier published version, or set a Step Functions alias to send all executions to an earlier version, from the console, CLI or API, with no rebuild; CodeDeploy can also roll back automatically on an alarm.[7, 10, 13] Durable executions started through a version or alias keep the version they started on.[2] Each function and state machine is rolled back separately, so a multi-step workflow takes one rollback per resource. Avalanche Cloud rolls back every step of the workflow at once to any previous version in one click or one command, and running workflows keep their version.[25]
How is AWS Lambda billed?
By requests and duration in GB-seconds, with a free tier of one million requests and 400,000 GB-seconds a month; durable functions add charges for durable operations, data written and data retained. Step Functions Standard bills per state transition, including retries, with 4,000 free a month. As of October 27, 2026.[37, 38]
Can I see which step changed a document?
Yes, on Avalanche Cloud. Every run gets a versioned workspace, and each step's changes are committed as a new version, so you can see which step produced or changed each file, compare versions and download any step's files.[25] On Lambda, files go in temporary /tmp storage, EFS or S3, and tracking which step changed a file is code you write.[32, 33]
06 / SourcesChecked 2026-10-27

Public primary sources.

Every source was fetched on the date shown. Prices and plans change; check the vendor before you buy.

  1. [1]Lambda durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-functions.html
  2. [2]Invoking Lambda durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-invoking.html
  3. [3]Basic concepts of durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-basic-concepts.html
  4. [4]Configure Lambda durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-configuration.html
  5. [5]Lambda quotas (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/gettingstarted-limits.html
  6. [6]Manage Lambda function versions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/configuration-versions.html
  7. [7]Create an alias for a Lambda function (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/configuration-aliases.html
  8. [8]Implement Lambda canary deployments using a weighted alias (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/configuring-alias-routing.html
  9. [9]Deployment configurations, AWS Lambda compute platform (AWS CodeDeploy User Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/codedeploy/latest/userguide/deployment-configurations.html
  10. [10]Redeploy and roll back a deployment with CodeDeploy (AWS CodeDeploy User Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/codedeploy/latest/userguide/deployments-rollback-and-redeploy.html
  11. [11]State machine versions in Step Functions workflows (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/concepts-state-machine-version.html
  12. [12]State machine aliases in Step Functions workflows (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/concepts-state-machine-alias.html
  13. [13]Example: Alias and version deployment in Step Functions (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/example-alias-version-deployment.html
  14. [14]How Step Functions associates executions with a version or alias (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/execution-alias-version-associate.html
  15. [15]UpdateStateMachine (AWS Step Functions API Reference)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/apireference/API_UpdateStateMachine.html
  16. [16]Step Functions service quotas (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/service-quotas.html
  17. [17]Restarting state machine executions with redrive (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/redrive-executions.html
  18. [18]Durable functions or Step Functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-step-functions.html
  19. [19]Monitoring durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-monitoring.html
  20. [20]Avalanche repository and README (MIT License)Trampoline AI · Checked https://github.com/Trampoline-AI/avalanche
  21. [21]Results, cancellation, and reruns (Avalanche docs)Avalanche docs · Checked https://avalanche-docs.docsalot.dev/build-workflows/results-and-reruns
  22. [22]Build a workflow with the Avalanche skill (Avalanche docs)Avalanche docs · Checked https://avalanche-docs.docsalot.dev/build-workflows/build-with-skill
  23. [23]Skills for agent steps: PDF, DOCX and spreadsheet (Avalanche docs)Avalanche docs · Checked https://avalanche-docs.docsalot.dev/agent-steps/skills
  24. [24]Architecture and current limits (Avalanche docs)Avalanche docs · Checked https://avalanche-docs.docsalot.dev/reference/architecture-and-limits
  25. [25]Avalanche and Avalanche Cloud overviewAvalanche · Checked https://avalanche.run/
  26. [26]Pull request #124: Claude Agent SDK agent stepsTrampoline AI on GitHub · Checked https://github.com/Trampoline-AI/avalanche/pull/124
  27. [27]Pull request #136: Agent Skills (SKILL.md) in agent stepsTrampoline AI on GitHub · Checked https://github.com/Trampoline-AI/avalanche/pull/136
  28. [28]Avalanche Cloud waitlistAvalanche · Checked https://cloud.avalanche.run/waitlist
  29. [29]Best practices for Lambda durable functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/durable-best-practices.html
  30. [30]Invoke a Lambda function on a schedule (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/with-eventbridge-scheduler.html
  31. [31]Creating and managing Lambda function URLs (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/urls-configuration.html
  32. [32]Configure ephemeral storage for Lambda functions (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/configuration-ephemeral-storage.html
  33. [33]Configuring Amazon EFS file system access (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/configuration-filesystem-efs.html
  34. [34]Identity and Access Management for AWS Lambda (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/security-iam.html
  35. [35]Logging AWS Lambda API calls using AWS CloudTrail (AWS Lambda Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/lambda/latest/dg/logging-using-cloudtrail.html
  36. [36]Managing your costs with AWS Budgets (AWS Cost Management User Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/cost-management/latest/userguide/budgets-managing-costs.html
  37. [37]AWS Lambda pricingAmazon Web Services · Checked https://aws.amazon.com/lambda/pricing/
  38. [38]AWS Step Functions pricingAmazon Web Services · Checked https://aws.amazon.com/step-functions/pricing/
  39. [39]Deploying a workflow that waits for human approval in Step Functions (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/tutorial-human-approval.html
  40. [40]Choosing workflow type in Step Functions (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/choosing-workflow-type.html
  41. [41]Invoke an AWS Lambda function with Step Functions (AWS Step Functions Developer Guide)Amazon Web Services · Checked https://docs.aws.amazon.com/step-functions/latest/dg/connect-lambda.html
Related comparisonsView all comparisons
See the runtime for yourselfOpen source · MIT

Our code speaks for itself.

The Avalanche runtime is open source under the MIT License. Read the code, run the starter project, and see how it fits the work your team repeats.