ConfluyeOpen app

RAG · Chat · Workflows · Claude Code · Codex

Confluye

The agent workspace where every tool shares context.

Upload the knowledge, ask with cited evidence, automate the workflow, and hand repository work to coding agents — on one graph that keeps the execution context.

Enter workspaceSee the demoWatch the film (45 seconds)
Acme workspaceShowing template: GitHub PR Code Review

Prompt

Help me set up the GitHub PR Code Review workflow.

Node 6 of 6 · Output

succeeded

Output

Example output · status: success · summary + 3 inline findings

Confluye in 45 seconds

Watch every tool flow into one graph.

Docs, chat, scripts, and coding agents flow into one graph. A pull request is reviewed by Codex and Claude Code with shared context, and the findings are published back to it.

Interactive product tour

See the whole system move as one.

Each section mirrors a surface teams actually operate. Choose one and the prompt, evidence, metrics, and workflow path change with it — nothing rotates on its own.

Managed knowledge

Documents become searchable context without user model setup.

Upload PDFs, docs, spreadsheets, or notes. Confluye indexes them with provider-managed embeddings, tracks reindex state, and exposes cited context to chat and workflows.

Indexed
248 docs
Reindex
Live
Sources
3 cited

Indexing and retrieving context

Which onboarding risk is highest in the uploaded Q3 support brief?

The highest risk is stalled enterprise onboarding: the support brief ties it to missing SSO docs, delayed security review, and repeated launch-date slips.

  • Q3 support brief.pdf
  • Security checklist.docx
  • Launch notes table
  1. step 1Upload docsPDF, Sheets, Drive
  2. step 2Chunk + embedProvider-managed embeddings
  3. step 3Answer with evidenceCitations in chat

What Confluye replaces

One operational layer for docs, chat, workflows, and runs.

  1. 01

    Knowledge-native context

    Knowledge bases, documents, and tables stay addressable from chat and from every workflow node, with citations that point back to the source.

    • Managed embeddings
    • Reindex state
    • Cited answers
  2. 02

    Agent workflow canvas

    Codex, Claude Code, MCP tools, conditions, webhooks, and approvals run as connected steps instead of disconnected scripts.

    • Triggers
    • Tool nodes
    • Approvals
  3. 03

    Provider-managed AI

    Embedding infrastructure stays invisible to customers while BYOK remains available for chat models, so teams start without provisioning anything.

    • No model setup
    • BYOK chat
    • Managed indexing
  4. 04

    Execution you can audit

    Runs keep their logs, evidence, and checkpoints attached to the workspace, so a review answers what happened without leaving the product.

    • Run logs
    • Evidence trail
    • Checkpoints

Context that carries over

Agents that keep context and run on a schedule.

Illustrative scenario: a real-estate agent uses an assistant inside a CRM built on Confluye. The assistant serves the agent, so the agent's own preferences, work in progress, and portal login stay with them. Buyers are client records the agent manages and chooses between. Each guide below says what you need to turn the feature on.

  • Memory that stays with the right person

    Your private memory

    Your agent can carry preferences, instructions, and facts across chats in one workspace. It saves durable memory only with your confirmation, and other members and admins cannot read it.

    Read about private agent memory

    Your customers' memory

    A product you build on Confluye can serve many customers from one workflow. Each customer gets their own conversation, facts, and work in progress. The two kinds of memory never mix.

    Learn about memory for your customers

  • Website logins your agent can reuse

    Your logins

    Sign in to a website yourself and save the login. You then decide which browser, Codex, or Claude Code blocks may use it. The login stays in Confluye's browser and never appears in the conversation.

    Set up saved website logins

    Your customers' logins

    Each customer signs in to their own website from a one-time link, and Confluye keeps every customer's login separate. Logins can expire; when one does, the customer gets a new link and signs in again.

    Customer website logins

  • Automations you can manage

    Scheduled workflows

    Schedule a deployed workflow, then pause, resume, run now, or edit it, and review its runs in the logs. A scheduled workflow repeats the same steps each time.

    Manage scheduled tasks

    Adaptive jobs

    When the plan itself should change, an adaptive job re-plans a Command Center prompt with AI on every recurrence. You choose one mode per automation.

    Compare the two modes

    People in the loop

    Add a Human in the Loop block where a person must decide. The run pauses, for minutes or days, until someone approves, rejects, or picks another outcome in the Tasks inbox; an expired task never counts as approval. A Wait block pauses for a set time or until another system calls back.

    Add a human decision

Availability depends on your deployment and permissions: memory and customer features stay off until an operator enables them, and workspace roles decide who can set them up.

Build on your agent runtime

One graph. Every agent. Full execution context.

Start in Confluye

Pricing

Pay for active compute, not for waiting.

Every plan includes the full workspace: knowledge bases, chat, workflows, and coding agents. Start with a 14-day trial with 2,000 credits and no card required.

  • Starter

    For one team automating its first production workflows.

    MXN 699/ month + IVA

    • 20,000 compute credits / month
    • 2 concurrent runs
    • 25 active workflows
    • 5 users
    • 1 workspace
    • 14-day detailed run history
    Get started

  • Pro

    For teams running several workflows and agents every day.

    MXN 1,999/ month + IVA

    • 100,000 compute credits / month
    • 5 concurrent runs
    • Unlimited active workflows
    • 15 users
    • 2 workspaces
    • 30-day detailed run history
    Get started

  • Team

    For organizations operating across several workspaces.

    MXN 4,900/ month + IVA

    • 400,000 compute credits / month
    • 15 concurrent runs
    • Unlimited active workflows
    • Unlimited users
    • 10 workspaces
    • 90-day detailed run history
    Get started

One credit covers up to 30 seconds of active standard compute; heavy runs use 4×. Time waiting for approvals, timers, or events is free. Extra usage is optional at MXN 0.03 per credit and never exceeds the monthly cap you set. Your own AI provider keys are billed by your provider.

Need more? Talk to us

Confluye Enterprise

Automation, governance, and a team to help you launch.

Bring your workflows into production with onboarding, advanced support, and a defined allowance of expert hours included in your software subscription.

Custom pricing

We agree on compute credits, onboarding scope, ongoing hours, and support coverage in your proposal.

Request a proposalExplore enterprise controls
Onboarding with a clear scope
Plan your first use case, configure existing connections, and train your team. Initial onboarding and recurring expert hours are defined separately.
Ongoing support and guidance
Get help with platform issues and use your expert hours for workflow reviews, training, and scoped improvements. Support channels and response targets are agreed up front.
Identity and change governance
Configure organization SSO and provisioning, workflow version approvals, and audit trails around how your team works.
Compute you can account for
Agree on execution credits and review usage across workspaces and workflows. Your AI provider subscriptions and BYOK usage remain separate.

Let’s talk about your team.

Tell us what you want to build. We’ll help scope a plan with the right compute, onboarding, and expert hours.

All fields are required unless marked optional.

Team size (optional)
A short overview is enough. Please leave out passwords and confidential data.

We’ll use these details to respond to your request. Read our Privacy Policy.

Prefer email? support@confluye.run

Google connections

How Confluye uses Google data

Google Sign-In provides your name and verified email so you can create or access your Confluye account. If you separately connect Google Drive, Docs, Sheets, or Gmail, Confluye accesses only the data and permissions you approve to retrieve content and run the workflows you configure.

Confluye uses Google user data only to provide the features you request. It does not sell Google user data or use it for advertising. Read the full details in our Privacy Policy.