# Alfrig

> Ask Alfrig in your own words: AI agents fix and test, your engineers approve, and you tell the customer. Start free, no card needed.

Alfrig is a team of AI agents that fixes and improves a company's software, with Alfrig as the way in. Describe the problem in your own words, with a screenshot or a screen recording. AI agents investigate, fix and test, your engineers approve what goes live, and you tell the customer at the right moment. By default, nothing goes straight to production: fixes land on the branch your engineers choose, after review by another agent and green CI. Web app at https://app.alfrig.com/ and desktop app for macOS, Windows and Linux. Site in English, Portuguese and Spanish.

# Who it's for

## Support and CS

The customer asks. Now you know the answer. Ask for the fix in the customer’s words, follow it on a timeline anyone understands, and get the reply ready when it’s live. No specs, no PRs, no engineer needed to translate.

**Before:** You open a ticket, wait for someone technical to get it, ask in the channel whether it’s done, and tell the customer “we’re looking into it”.

**With Alfrig:** You explain the problem, answer one or two questions, follow the timeline, and reply to the customer the day it goes live.

### Ask the way you talk

Type or speak, attach a screenshot or a screen recording, or paste the ticket link. Alfrig shapes the request for you.

### Answer with one click

When something is missing, the question is simple and comes with options: one click in the Alfrig alert, or the Answer button in Slack.

### Follow along without chasing status

In Progress, the My requests filter shows just yours: received, understood, fixed in test, released, live, with when each one is expected.

### Tell the customer at the right time

When it reaches production, you get an alert and a draft reply.

You see your requests, in business language. Specs, code and repository names stay with your engineers.

### Do I need to know anything technical?

No. Say what the customer saw and what should have happened. Alfrig asks the rest, with options to click.

### What if I explain it wrong?

Alfrig confirms what it understood when in doubt, and a screenshot or screen recording settles most cases. If the delivery isn’t right, you ask for the change on the request itself.

### Can I break something?

No. Every fix is checked by another agent and by the tests, and only goes live the way your engineers set it up.

### Does it work from Slack?

Alerts arrive in Slack with buttons to reply and approve. Opening a request is done in the chat with Alfrig, with a label in your tracker or through the tracking link.

### Can people without an account follow along?

Yes. Every request has a link with its timeline that opens without logging in.

## CEO and COO

Your company responds faster, without growing the team. The people who talk to customers ask, AI agents deliver reviewed fixes and improvements, and you see what’s moving, what needs someone and what was delivered, without asking anyone.

**Before:** The small queue competes with the roadmap, and nobody knows what’s moving without asking.

**With Alfrig:** The queue moves in parallel, engineering stays on what matters, and progress and the month’s numbers are in plain sight.

### Capacity that actually delivers

Each delivery is a fix or improvement that’s done, checked by another agent and by the tests, and approved by engineering.

### Everything in plain sight

In Progress, who is working, what needs someone and what was delivered. In Performance, the month’s tasks, what QA sent back, time to answer and cost.

### Ideas with an owner until the end

An idea becomes a proposal, screens and a delivery, with an agent in charge that only stops when it needs someone.

### Pay per delivery, no surprises

You pay for what gets done, in a monthly plan with an overage cap. Unlimited members.

You see progress on everything and the numbers that matter, without opening a PR or reading code.

### What does it really cost?

The trial is free, with 5 deliveries. After that, a monthly plan by deliveries: Team is US$ 249 a month with the AI included (or US$ 99 with your company’s AI account), extra deliveries stay within the spending limit you set, and people are unlimited. The Pricing calculator shows the cost for your volume.

### Will engineering accept it?

They stay in control: they decide where fixes go, review whatever they want and choose how they reach production. And they stop translating poorly described requests.

### Is our data safe?

Each organization has its own database and encrypted keys. The agent works in an isolated, disposable environment and never sees the repository credentials.

### How soon will I see results?

Engineering connects the code in a few minutes, and Alfrig already suggests the first delivery from the mapping of your systems.

## Engineering

The request queue moves. So does the roadmap. Requests that arrive with repro steps, evidence and a definition of done. One agent writes the spec and fixes, another reviews against every criterion, CI passes, and the merge only lands on the branch you chose. By default, nothing reaches production without your team’s decision.

**Before:** Support writes “the discount is wrong”, you drop what you’re doing, reproduce, fix, open the PR and explain in the channel when it’ll go live.

**With Alfrig:** The request arrives with repro steps and a definition of done, one agent fixes and another reviews. You check the PR when you want and decide what goes to production.

### Connect in minutes

GitHub app (or a token for GitLab, Bitbucket and Azure Repos). Alfrig maps the catalog: stack, test command and dependencies between projects.

### Specs with verifiable criteria

Each acceptance criterion says how it’s checked: test, QA or manual. The review checks them one by one.

### Review by another agent and CI

The author never approves its own PR. Merge only with green CI, in dependency order, on the agreed branch.

### Several repositories in one task

Backend and frontend in the same task, picked from the catalog, with one agent per repository when agents are free and merges in dependency order.

You see everything: spec, diffs, commands, tests, time and cost for every step, and you decide what needs approval.

### Is the code any good?

Every task starts with a spec and acceptance criteria. Tests related to the change run before the PR, and another agent reviews against each criterion, in fix rounds up to a limit (3 by default). If tests fail, it becomes a draft PR, never a merge.

### What access does Alfrig need?

Only to the repositories you choose: through the Alfrig app on GitHub or a token on other providers. Access is encrypted and never handed to the agent.

### Does the agent see our credentials?

No. Clone, push, PR and merge are done by Alfrig; the agent works on a copy with no token, in an isolated environment with no inbound connections, destroyed at the end.

### Is our code used to train models?

Alfrig does not train models. Data use depends on the model provider you connect.

### Does it work with our stack?

Node, PHP, Python, Java/Kotlin, .NET, Go, Ruby and Flutter, plus mixed projects, monorepos included.

## Product

From a loose idea to delivery, with one owner from start to finish. Tell Alfrig the idea. An agent in charge understands the problem, proposes the deliveries, designs the screens with your product’s design system and follows the build until the last delivery. It only stops for questions, the proposal and the screens.

**Before:** The idea stays in a conversation; nobody writes the proposal, designs the screens or follows it to the end.

**With Alfrig:** You share the idea, answer the questions that change the decision, approve the proposal and the screens, and watch the deliveries arrive.

### Understand before proposing

The problem, who feels it and what already exists. If needed, the agent reads the code before concluding.

### A proposal you can edit

The problem, who feels it, what changes and how you’ll know it worked, the deliveries in order and what’s left out. You untick what you don’t want now, approve or ask for changes.

### Screens that look like your product

Designed on a board with the design system surveyed from your code, checked in a browser.

### A build followed to the end

Deliveries become tasks in batches, in dependency order, and the initiative only ends when all of them are done.

You see the proposal, the screens and every delivery moving; engineering reviews the code before it goes live.

### Will the screens look like my product?

Yes. The agent surveys your frontend’s tokens, patterns and components from the code, and your brand rules apply to every screen.

### Can I change the proposal?

Yes. Untick deliveries, ask for changes in a sentence, or approve it as it is.

### Who builds it afterwards?

The team’s agents, one delivery at a time, with review and tests like any request. The owner follows it to the last one.

### What if the idea is still rough?

That’s what the understand step is for: the owner asks what changes the decision before proposing.

# How it works

Whoever knows the problem asks, agents resolve it with review, everyone follows along and the customer hears at the right time. Production follows your engineers’ call.

## 1. Set up: Engineering connects, Alfrig works out the rest.

A guided walkthrough that takes a few minutes. The Ready for agents panel shows what’s missing for the first fix, and Alfrig works out almost everything on its own.

- Connect the code: Alfrig reads each repository and writes its card (what it is, the stack, how it’s tested and what it depends on).
- Where fixes go: agents deliver to a test branch; your team takes it to production.
- Requests and notices: a label in ClickUp, Jira, Linear or Azure Boards opens the task, and notices go to Slack, Teams or email.
- Invite the people who ask: CS and Product open requests, answer and follow along, without seeing settings.
- Check in the test environment: after the merge, QA checks each delivery the way a person would.
- The first delivery: a small, real task Alfrig spotted in your code.

## 2. Ask: In your own words, the easiest way.

In the chat with Alfrig, type or speak, attach a screenshot or a screen recording, or paste the ticket link. Alfrig understands the problem, asks only what’s missing and shapes the request.

- Text, voice, a screenshot, a screen recording or the ticket link.
- Alfrig asks what’s missing, with options to click: what happens, what should happen, how to reproduce it and where.
- A broad idea becomes a Product initiative, with an agent in charge from start to finish.
- Also through a label in your tracker, the API or an AI assistant over MCP.

## 3. Resolve: One agent fixes, another checks.

Every request becomes a spec with acceptance criteria. One agent investigates, fixes and tests; another reviews criterion by criterion; with green CI, the merge lands on the agreed branch.

- Each criterion says how it’s checked: test, QA or manual.
- When confidence is low, Alfrig confirms what it understood with whoever asked.
- Review in rounds up to a limit (3 by default); if tests fail, it becomes a draft PR, never a merge.
- Several repositories in one task, merged in dependency order.
- Every task leaves lessons about the project and keeps AGENTS.md up to date through PRs.

## 4. Follow: Everything in plain sight, without asking anyone.

Progress shows who is working, what needs you and what was delivered. Notices arrive with the answer right there, in the app, Slack or Teams.

- Questions with options, answered with one click in the notice.
- Ask Alfrig "how is it going?" or ask it to approve, resume or comment on the ticket: each action becomes a card that only runs once you confirm.
- Each agent is a mascot that shows whether it’s free, working, needs you or has the delivery ready.
- The My requests filter shows only what you asked for.

## 5. Tell the customer: You reply to the customer the day it goes live.

The timeline runs from received to live. When it reaches production, you get a notice and a suggested reply to review before sending.

- Received, understood, fixed in test, released, live.
- Released means it goes to production with the next release; live means it’s already there.
- A suggested reply to the customer, ready to copy.
- A tracking link that opens without signing in.

### How long does it take to get started?

Someone from engineering connects the code and confirms where fixes go in a few minutes. Alfrig maps the repositories and suggests the first delivery right away.

### How does Alfrig know which repository to change?

When you connect, an agent reads each project, writes its card (what it does, stack and how to test it) and finds the dependencies between them. Agents use that map to choose where to change and in which order to merge.

### Does Alfrig do more than fix things?

Yes. A broad request becomes a Product initiative: an agent in charge understands the problem, proposes the deliveries, designs the screens and follows the build. And there’s Strategy: OKRs built in a chat that set the priority of initiatives.

### How do the agents learn?

When a task ends, the agent saves short lessons about the project that go into the context of the next ones; you review, write and delete whatever you want. The project’s AGENTS.md is written when missing and updated through PRs.

### Who decides what goes to production?

Your engineers. By default, fixes land on the agreed branch after review and CI; taking them to production is your team’s call.

# Security

By default, nothing goes straight to production: fixes land on the branch your engineers choose, after review and CI. And the agent works in isolation, without your repository credentials.

## Merge safeguards: What an agent produces goes through more than one door.

Auto-merge is optional and starts off. Even when it’s on, it waits for a person when something is out of the ordinary.

- **Green CI, always**: Merges only happen with an approved review and green CI. Red CI in any repository of the task holds them all.
- **Pipeline and code owners protected**: An agent’s changes to CI/CD and CODEOWNERS are left out of the push by default; when the connection allows them, the PR waits for a person.
- **Build and data collection under review**: Changes to build and config files (Makefile, Dockerfile, .npmrc and the like) or new data-collection points hold auto-merge.
- **Outside text gives no orders**: What comes from tickets, comments and pages is treated as information, never as instructions. If the agent raises an alert, auto-merge turns off.
- **Tests don’t vanish to pass**: Deleted or skipped tests are flagged in review. If tests fail, it becomes a draft PR, never a merge.

## Isolation: One machine per step, without your credentials.

Each step runs on a machine created for it and destroyed when it ends.

- **Disposable machine**: No inbound connections, destroyed at the end of the step; a collector removes any left behind.
- **Tests run offline**: Tests run with the network turned off.
- **Alfrig keeps the token**: Clone, push, PR and merge are done by Alfrig. The agent works without the repository token, which lives only in memory, never in the repository URL or on disk.

## Data: Each organization with its own database and its own key.

One organization’s data never shares a database with another’s.

- **Your own database**: Every organization has its own database.
- **Encrypted keys**: Tokens and keys are encrypted with AES-256-GCM, with a key unique to each organization.
- **Signed integrations**: Public API keys are stored only as hashes. Webhooks Alfrig sends are signed, and the ones it receives are verified.
- **Where it runs**: The application runs in São Paulo, Brazil. Agent machines do too, or in the US when capacity there runs out.

## People and access: Each person reaches what their role allows.

Invite the people who ask without opening the settings to everyone.

- **Five roles**: Owner, admin, member, requester and guest. Requesters see only their own requests; workflow, models, costs and keys stay with admins and the owner.
- **Connected apps**: An AI assistant connected over MCP reaches only what the authorizing person’s role allows, checked on every call.
- **Sign-in**: With GitHub, Google, Apple or an email link.

## AI models: Alfrig doesn’t train models on your code.

You choose whose AI account it is.

- **No training**: Alfrig doesn’t train models. Data use follows the policy of the provider handling the call.
- **With your key**: Anthropic, OpenAI, Google Gemini or z.ai: the model is paid to the provider directly. Calls go through the Alfrig gateway, which uses your key and meters each step, with no markup.

## Deletion and privacy: Delete means delete.

The organization owner decides when the data stops existing.

- **Real deletion**: The owner can delete the organization for good: database, files and the encryption key, which leaves any remaining backup unreadable.
- **Short retention**: Technical logs for up to 2 months; backups removed within 30 days after deletion.
- **Your rights**: Access, correction, portability and deletion, through the contact in the privacy policy.

### Does Alfrig release to production on its own?

Not by default. The organization starts with auto-merge off, and fixes go to the branch your engineers choose. If your team turns auto-merge on with the production branch as the target, the fix reaches production with no one stepping in: that’s their call.

### Does the agent see our credentials?

Not the repository ones. Clone, push, PR and merge are done by Alfrig, and the agent works on its own machine, without the token, destroyed at the end of the step.

### What access does Alfrig need?

Only to the repositories you choose: through the Alfrig app on GitHub or a token on other providers. Access is encrypted with your organization’s key.

### Does our code train models?

Alfrig doesn’t train models. Data use depends on the model provider handling the call.

### Where is the data?

In a database that belongs only to your organization. The application runs in São Paulo, Brazil; agent machines do too, or in the US when capacity there runs out.

# Pricing

You pay for what gets done: each delivery is a fix or improvement that’s done, checked by another agent and by the tests, and approved. Unlimited people on every plan.

- **Trial** — US$ 0 (All included: US$ 0). See your first deliveries before you decide. 5 deliveries on us; Valid for 14 days; No card and nothing to set up; Unlimited people.
- **Team** — US$ 99 /month (All included: US$ 249). To take the request queue off your team’s plate. 40 deliveries a month; US$ 3 per extra delivery; You set how much extra you can spend; Unlimited people; Email support.
- **Scale** — US$ 299 /month (All included: US$ 799). For several products and a queue that never hits zero. 150 deliveries a month; US$ 2 per extra delivery; You set how much extra you can spend; Unlimited people; Email support.
- **Enterprise** — Custom (All included: Custom). For multi-product groups and IT requirements. Delivery volume to fit; AI included or your own account; Dedicated infrastructure; Single sign-on (SSO) and access policies; Invoicing, contract, SLA and priority support.

# FAQ

### What is Alfrig?

A team of AI agents that fixes and improves your company’s software, with Alfrig as the way in. You ask in a chat, in your own words; one agent investigates, fixes and tests; another checks it; your engineers approve what goes live; and you get an alert to reply to the customer.

### Do I need to know how to code?

No. You explain the problem like you would to a colleague, by text, voice, screenshot or screen recording, or paste the ticket link. Alfrig asks what’s missing with options to click and shows progress in business language.

### Who decides what goes live?

Your engineers. By default, Alfrig doesn’t ship to production: fixes land on the agreed branch, after being checked by another agent and by the tests.

### How does engineering get involved?

Someone on the team connects the code (through the GitHub app, or GitLab, Bitbucket and Azure DevOps) and confirms where fixes go. It takes a few minutes. If you’re not technical, use the “Send to my engineers” button.

### How do I know when to tell the customer?

Every request has a timeline: received, understood, fixed in testing, released, live and customer told. When it reaches production, you get an alert and a draft reply for the customer.

### Where can I ask from?

From the chat with Alfrig, a label in ClickUp, Jira, Linear or Azure Boards, the API, or an AI assistant through MCP. Alerts arrive in Slack, Teams and email, with buttons to reply and approve.

### What if Alfrig gets it wrong?

When confidence is low, it confirms what it understood with you before starting. And if the delivery isn’t right, you ask for the change on the request itself.

### How much does it cost?

The trial is free: 5 deliveries on us, no card. After that, you pay per delivery in a monthly plan, with unlimited members. Details are under Pricing.

## Links

- [Home (English)](https://alfrig.com/en/)
- [Support and CS (English)](https://alfrig.com/en/for/support/)
- [CEO and COO (English)](https://alfrig.com/en/for/leadership/)
- [Engineering (English)](https://alfrig.com/en/for/engineering/)
- [Product (English)](https://alfrig.com/en/for/product/)
- [How it works (English)](https://alfrig.com/en/how-it-works/)
- [Security (English)](https://alfrig.com/en/security/)
- [Home (Português)](https://alfrig.com/)
- [Suporte e CS (Português)](https://alfrig.com/para/atendimento/)
- [CEO e COO (Português)](https://alfrig.com/para/gestao/)
- [Engenharia (Português)](https://alfrig.com/para/engenharia/)
- [Produto (Português)](https://alfrig.com/para/produto/)
- [Como funciona (Português)](https://alfrig.com/como-funciona/)
- [Segurança (Português)](https://alfrig.com/seguranca/)
- [Home (Español)](https://alfrig.com/es/)
- [Soporte y CS (Español)](https://alfrig.com/es/para/soporte/)
- [CEO y COO (Español)](https://alfrig.com/es/para/direccion/)
- [Ingeniería (Español)](https://alfrig.com/es/para/ingenieria/)
- [Producto (Español)](https://alfrig.com/es/para/producto/)
- [Cómo funciona (Español)](https://alfrig.com/es/como-funciona/)
- [Seguridad (Español)](https://alfrig.com/es/seguridad/)
- [Start free](https://app.alfrig.com/)
- [Product Hunt](https://www.producthunt.com/products/alfrig-ai-agents-pull-request) · [YouTube](https://www.youtube.com/channel/UCuWgwhpUp2h3XgMaTQT23RA)
- Contact: contato@alfrig.com