# Zencoder — AI coding agent platform with desktop, IDE and CI agents that plan, build, test and review across repositories.

> Source: The Agents Index — https://theagentsindex.com/zencoder (structured, researched, re-verified)
> Facts last verified: 2026-09-09

Zencoder is an AI coding agent platform built around Zenflow, a desktop app that runs agents in isolated Git worktrees, plus plugins for VS Code, JetBrains and Android Studio. A second app, Zenflow Work, points the same agents at business tools.

| Fact | Value |
| --- | --- |
| Website | https://zencoder.ai |
| Price | from $45/month billed monthly, $40/month billed annually |
| Free tier | Yes |

A standing Free plan is priced at $0 but its credit allowance is listed as "BYOK only" with models limited to "Open / Auto" (documented elsewhere as Auto, Gemini Flash 3.0 and Grok Code Fast 1), so sustained use requires your own OpenAI or Anthropic key. Separately, a 7-day Pro trial grants 5,000 credits and ends at 7 days or 5,000 credits, whichever comes first; leftover trial credits do not carry over. Cancelled paid accounts drop back to Free.

## What it does

**Writes code** · Yes — Generates and modifies code across many files and languages from one prompt, then validates its own output and self-corrects on lint or test failures.

> The Coding Agent is Zencoder's primary development tool that generates, modifies, and refactors code across multiple files and languages.
> — https://docs.zencoder.ai/features/coding-agent · What is the Coding Agent? section, opening paragraph

**Debugs and fixes** · Yes — A dedicated Fix a Bug workflow investigates root cause, writes a solution document for review, then implements the fix and confirms the regression is gone.

> The agent implements the fix, runs tests, and verifies the regression is resolved. Results go into implementation.md.
> — https://docs.zencoder.ai/zenflow/workflows/fix-a-bug · How It Works, Implementation step

**Writes tests** · Yes — Writing tests is named as a first-class task type, and after each task Zenflow runs the repository's own linters, type-checkers and test suites as verification.

> Zenflow Code works directly within your codebase to execute complex multi-file engineering tasks — such as implementing large features, refactoring components, fixing bugs, and writing tests.
> — https://docs.zencoder.ai/zenflow/overview · What Zenflow Code Is, first paragraph

**Reviews code** · Yes — Review is a separate workflow step that can be assigned a different model than the one that wrote the code, and agents can review incoming pull requests.

> Automated code reviews | Agent reviews incoming pull requests and posts feedback
> — https://docs.zencoder.ai/admin/agents · Use Cases table, Automated code reviews row

**Whole-codebase work** · Partly — Agents work across a whole repository on any plan, but the cross-repository indexing that carries a change through several services starts at Pro Plus and is admin-configured.

> Multi-Repository Search is available for users on Pro Plus and Pro Max plans. Repository and Connections management requires Owner or Manager role permissions.
> — https://docs.zencoder.ai/features/multi-repo · What is Multi-Repository Search? section, Tip callout

**Ships it** · Partly — It commits, pushes the task branch and opens a pull request with a pre-filled description, and can auto-push, but no page claims it deploys to an environment.

> Zenflow pushes the task branch and opens the PR on GitHub.
> — https://docs.zencoder.ai/zenflow/integrations/github · Features, Create Pull Requests subsection

## How much it does unattended

**Runs autonomously** · Yes — Agents complete tasks unattended on a schedule or from a repository event, load context, edit, run tests and publish results without a person approving each step.

> deployed as headless workers that live in your CI/CD systems... they load the repo map, follow their instructions, plan edits, run tests, and publish output without a human steering every step.
> — https://docs.zencoder.ai/learn/enterprise-best-practices/ci-cd-automation · GitHub-Triggered Autonomous Agents card, What are autonomous agents?

**Multi-agent** · Yes — Several agents run at once in separate Git worktrees, and a single workflow can hand planning, implementation and review to different agents and models.

> Zenflow's parallel execution model lets you run multiple agents simultaneously across different tasks.
> — https://docs.zencoder.ai/zenflow/blast-agents · Overview section, first paragraph

**Agent permissions** · Yes — Every saved preset carries a permission mode, MCP tool calls prompt per invocation unless marked always-allow, and committing or pushing is a separate opt-in setting.

> Agents request explicit permission before invoking MCP tools:
> — https://docs.zencoder.ai/features/security · MCP Tool Security section

## Which models it runs on

**Claude** · Yes — Anthropic models are in the managed catalogue on paid plans and Claude Code can be chosen as the runtime; the Opus tiers require Pro Plus or above.

> Anthropic | Haiku 4.5, Sonnet 4.6, Opus 4.6, Opus 4.7
> — https://docs.zencoder.ai/clis/overview · Model Flexibility table, Anthropic row

**GPT** · Yes — OpenAI models are in the managed catalogue on paid plans and the Codex CLI can run as the agent runtime; the largest GPT tier needs Pro Plus or above.

> OpenAI | GPT-5.3 Codex, GPT-5.4, GPT-5.4-mini, GPT-5.5
> — https://docs.zencoder.ai/clis/overview · Model Flexibility table, OpenAI row

**Open models** · Partly — Open-weight models are not sold in the managed catalogue; you reach them by declaring your own Ollama or vLLM endpoint in settings.json and paying for that inference yourself.

> Local models: Run open-source models via Ollama or vLLM on your machine
> — https://docs.zencoder.ai/clis/custom-agent · Use Cases list, Local models bullet

**Model choice** · Yes — You pick the model per chat and per workflow step from a published catalogue, and can switch mid-task; Auto routing is one option rather than the only path.

> Zencoder has a model selector in the chat input. Use the dropdown to pick which LLM runs your messages. The options you see depend on your plan and entitlements.
> — https://docs.zencoder.ai/features/models · Overview section, first paragraph

## Where you use it

**In your editor** · Yes — Marketplace plugins put the agent inside VS Code, JetBrains IDEs and Android Studio, where diffs are reviewed and accepted without leaving the editor.

> Zencoder's IDE plugins provide standard AI co-pilot functions and coding agents directly inside your favorite editor.
> — https://docs.zencoder.ai/get-started/install-ide-plugin · Install IDE Plugin, opening line

**On the command line** · Partly — Terminal agents are supported as runtimes — its own Zen CLI plus Claude Code and Codex — but the docs frame them as hosted inside the editor rather than as a standalone Zencoder terminal product.

> Zencoder acts as a host for external AI coding CLIs, giving them IDE integration, multi-repo context, custom agents, and skills. You pick the runtime; Zencoder provides the surrounding infrastructure.
> — https://docs.zencoder.ai/features/universal-cli-platform · What It Does section

**In your pipeline** · Partly — Headless agents can be triggered from GitHub, GitLab, Bitbucket or a webhook and act on pull requests, but that autonomous-flows capability is sold as a paid add-on from Pro Plus upward.

> It's a paid addon (Pro Plus plan or higher). Contact sales via
> — https://docs.zencoder.ai/learn/enterprise-best-practices/ci-cd-automation · GitHub-Triggered Autonomous Agents card, How can autonomous agents be activated at Zencoder, as an addon?

**In a browser** · Partly — The main app is a desktop download; the only documented browser route to it is installing Zenflow on your own VPS, while auth.zencoder.ai covers administration and analytics rather than agent work.

> Once logged in, you'll see the onboarding screen of Zenflow and now can work with it through browser remotely
> — https://docs.zencoder.ai/zenflow/vps-setup · Setup steps, step 8

## Whose machine it runs on

**Self-hosted** · Partly — Zenflow can be installed on a VPS you own, and a full private deployment running orchestration and inference inside your network is listed as an Enterprise-plan feature.

> A private deployment gives you deeper control by running agents, orchestration, and inference on infrastructure you operate.
> — https://docs.zencoder.ai/features/private-deployments · What Is a Private Deployment? section

_Not established: Open source_

## What it costs to run

**How it meters** · Yes — Billing pairs a per-seat monthly fee with metered credits, where each model carries a published multiplier and extra credits are bought on demand at a $20 minimum.

> Zencoder uses a usage-based credit model. Every paid plan includes a flat seat price plus a monthly credit allowance. Credits are consumed by LLM usage.
> — https://docs.zencoder.ai/faq/pricing · Pricing & Plans, opening paragraph

**Free tier** · Yes — A standing $0 Free plan exists but carries no bundled credits, so continuous use needs your own provider key; a 7-day Pro trial adds 5,000 credits once.

> You move to Free at the end of the current billing period. Plan credits are forfeited. Existing top-ups remain usable. After top-ups run out, your account is BYOK-only.
> — https://docs.zencoder.ai/faq/pricing · FAQ, What happens if I cancel? accordion

**API access** · Partly — Org API keys authenticate CI integrations and the VPS install, but the only REST endpoint documented is organisation usage analytics, gated to Pro Plus and an Owner or Manager role.

> API access is available for users on Pro Plus plans and above. Requires Owner or Manager role permissions within your organization.
> — https://docs.zencoder.ai/features/analytics-api · Overview section, Note callout

**MCP server** · No — Zencoder consumes MCP servers rather than publishing one, so another agent has no documented MCP endpoint through which to query or drive it.

> Zencoder acts as an MCP client — it connects to MCP servers that expose specific capabilities like database queries, API calls, or browser automation.
> — https://docs.zencoder.ai/features/mcp-deep-dive · What is MCP? section

**Bring your own key** · Yes — Your own OpenAI or Anthropic key works on every plan and those calls consume no credits, and custom providers go further with local, VPC or third-party endpoints.

> All plans — including Free — support BYOK. When you use your own API key (OpenAI, Anthropic, or Gemini), LLM calls go through your key and do not consume bundled credits.
> — https://docs.zencoder.ai/faq/pricing · BYOK (Bring Your Own Key) section

## Buying it for a team

**A company can buy it** · Yes — Seats are bought and managed centrally: an Owner controls billing and members, Managers handle users and repositories, and Enterprise terms are quoted by sales.

> Owner | Full organization control — billing, users, settings, repositories
> — https://docs.zencoder.ai/features/security · Role-Based Access Control table, Owner row

**Seat model** · Yes — Priced per user per month with a credit allowance attached to each seat; seats are freed and reassigned from the admin panel and no minimum seat count is published.

> Per-seat credits that refresh each month.
> — https://zencoder.ai/pricing · Pro plan card, tooltip on the 30,000 monthly credits bullet

**Pooled budget** · Partly — Credits pool across the organisation with admin-set per-user caps from Pro Plus upward, while the Pro tier is listed as per-seat credits only.

> Pooled across all seats with per-user caps.
> — https://zencoder.ai/pricing · Pro Plus plan card, tooltip on the Shared team credit pool bullet

**Admin controls** · Yes — Admins cap how much of the shared pool each user may burn, choose which repositories agents can index and read, and can restrict repository access to named emails.

> Admins can set a max-% cap of the org credit pool per user... A user who hits their cap is blocked from further usage until the admin re-allocates or buys a top-up.
> — https://docs.zencoder.ai/faq/pricing · Team & Admin Controls, Per-user credit caps accordion

**Audit log** · Partly — Authentication, user-management, configuration and usage events are recorded and exportable as CSV, but only from Pro Plus upward — the entry Pro tier has none.

> Available on Pro Plus, Pro Max, and Enterprise plans.
> — https://docs.zencoder.ai/features/team-administration · Audit Logs section, availability line

**Single sign-on** · Partly — SAML 2.0 and OIDC are supported with enforced SSO-only login and automatic provisioning, but the capability begins at Pro Plus rather than the entry paid tier.

> Available on Pro Plus, Pro Max, and Enterprise plans... SAML 2.0 and OIDC support
> — https://docs.zencoder.ai/features/security · Enterprise Security Controls section, availability line and SSO (Single Sign-On) bullet list

## What happens to your code

**Opt out of training** · Yes — Training on customer code is refused outright rather than offered as a toggle, and code context is described as processed ephemerally unless persistence is switched on.

> No model training on your code — Your code is never used to train or fine-tune any AI model
> — https://docs.zencoder.ai/features/security · What Zencoder Does NOT Do, first bullet

**Data residency** · Partly — No hosting region or jurisdiction is named for the managed cloud and no region picker is documented; the published answer to a residency mandate is to run a private deployment yourself.

> Private deployments are recommended for regulated industries, strict data-residency mandates, air-gapped environments, and teams with proprietary models they must keep on-premise.
> — https://docs.zencoder.ai/features/private-deployments · What Is a Private Deployment? section, closing paragraph

**Getting out** · Partly — Audit logs and analytics export as CSV and repository indexes are deletable on demand, but plan credits expire every billing period with no rollover and top-ups are non-refundable.

> No. Plan credits expire at the end of your monthly billing period. Top-up credits, however, remain usable.
> — https://zencoder.ai/pricing · FAQ, Do unused plan credits roll over? accordion

**Certifications** · Yes — SOC 2 Type II, ISO 27001 and ISO 42001 certification are claimed alongside GDPR compliance, with reports offered through a hosted trust centre.

> SOC 2 Type II, ISO 27001, and ISO 42001. Full audit trails, zero code storage, zero model training on your data.
> — https://zencoder.ai · Built for teams shipping at scale, Security & Compliance card

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

None recorded. Every fact above comes from the vendor and is not independently corroborated yet.

## What changed

- 2026-09-09 — Capabilities · Commercial terms · Free tier · Price

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Zencoder, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.