GoCodeo
Also known as: GoCodeo, gocodeo.com, GoCodeo AI
VS Code extension that builds, tests and deploys full stack apps from a prompt, with ASK, BUILD, TEST and MCP flows, a model selector across four providers, 100+ tools via customer defined MCP servers, and one click Supabase and Vercel setup.
GoCodeo is an AI coding agent that lives inside Visual Studio Code and turns a natural language prompt into a working full stack application. Rather than acting as autocomplete, it scaffolds a project from starter templates across more than twenty five frameworks, generates front end and back end code, wires up a database, and deploys. It is organized around four flows reachable by keyboard shortcut. ASK handles chat, file explanations, document and image queries and terminal debugging. BUILD handles project generation and deployment, TEST handles test generation and execution, and MCP connects external tools.
BUILD generates a complete project from a prompt, or from an attached image, expanding short input into a fuller functional and technical specification first. It integrates natively with Supabase to scaffold database tables, authentication flows and storage, and with Vercel for one click deployment including generated configs, environment variables and CI/CD setup. Git operations are handled in one go. The agent analyzes the repository, commits, pushes and raises a pull request, returning a link. Build History keeps multiple build sessions organized and switchable.
MCP is how integration breadth is achieved. Customers define their own tools in an mcp.json file naming endpoints, arguments and tokens, including internal APIs, and the vendor claims more than a hundred connectable tools with GitHub, Notion, Postgres, Stripe and Perplexity named. Community built MCP servers are browsable and installable from inside the extension, and the agent selects the relevant tool by intent, chaining multi step workflows such as reading a specification from Notion, generating the application, pushing to GitHub and deploying to Vercel.
TEST is the most developed part of the product. It generates unit tests covering positive, negative, edge and error cases from a class or function, with the framework, code style and number of tests per run configurable, cases saveable for reuse, a feedback loop that refines future generation, AI explanations of failures, and a Fix Test/Code control. Test execution covers Python and JavaScript. A Model Selector lets the developer choose the model behind the agent across OpenAI, Claude, Gemini and DeepSeek.
GoCodeo's own pages disagree on pricing. Its marketplace listing says GoCodeo is now completely free because customers bring their own API keys, while its pricing page sells Hobby free, Starter at nine dollars and Pro at nineteen dollars a month, with metered premium model requests and a ten dollar add on for two hundred and fifty more. GoCodeo names no security certification, SSO, RBAC or audit surface. The site carries Terms of Use but no privacy policy or DPA, linking instead to the privacy policies of OpenAI, Anthropic and Google. GoCodeo says it does not store or train on customer code and that generated code belongs to the customer.
Some details speak to how recently the product was updated. The extension's latest version on the VS Code Marketplace is 11.0.13, published 9 July 2025, the site footer reads 2025, and the models the pricing page names stop at Claude 4 Sonnet, Claude Sonnet 3.5, GPT-4o, o3-mini and DeepSeek-R1. Invocation is manual and limited to one surface, by keyboard shortcut inside VS Code, with no event triggers, schedules or inbound channels, and there is no IntelliJ or JetBrains distribution.
Vendor details
Canonical URL
https://www.gocodeo.com
Category
Coding agent
Subcategory
IDE-native full stack coding agent (VS Code)
Funding status
Independent, early stage. GoCodeo is a small startup; specific funding figures and founder names are not disclosed on any first-party page. The VS Code Marketplace listing reports 39,692 installs against 18 ratings, while the website claims 25,000+ engineers. Company logos shown on the site (PwC, Cognizant, Capgemini, Intuit, Cisco, Salesforce, Autodesk, Amazon, TCS, Rubrik, Freshworks) are presented as 'developers from' those companies, i.e. individual users, not enterprise customers, and should not be read as customer references.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Runs as a Visual Studio Code extension. MCP is the integration framework, with 100+ connectable tools claimed and GitHub, Notion, Postgres, Stripe and Perplexity named. Customers define their own tools in an mcp.json file including internal APIs, and a Browse Community MCP Servers option exposes prebuilt community integrations. Native integrations cover Supabase (schema, auth, real time persistence), Vercel (one click deploy with generated configs and environment variables) and Git (repo init, commits, pull request creation with a link). An @Web search powered by Perplexity fetches live documentation. A Model Selector spans OpenAI, Claude, Gemini and DeepSeek. There is no IntelliJ or JetBrains distribution.
In practice
You want to stand up a full stack app fast without leaving your editor. You prompt GoCodeo in VS Code, and the agent scaffolds the project, generates front and back end code, wires up Supabase, and deploys to Vercel.
You dread writing tests. GoCodeo's TEST flow generates unit tests covering positive, negative, edge and error cases from a class or function, runs them in the editor, explains failures, and lets you set the framework and number of tests, then save reusable cases.
You want your agent to reach the tools you already use. Through the Model Context Protocol GoCodeo connects to tools like GitHub, Notion, and Stripe, out of more than one hundred it claims, and you can define your own in an mcp.json file, including internal APIs.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
7.0 / 14 capabilities · 50%
| Integrations & Tool Calling | Full |
|---|---|
|
MCP is the integration framework. GoCodeo claims 100+ connectable tools and names GitHub, Notion, Postgres, Stripe and Perplexity. Customers define their own tools, including internal APIs, in an mcp.json file they edit, and a Browse Community MCP Servers option offers prebuilt integrations shared by other developers. The agent selects tools by intent. Native integrations cover Supabase for schema, authentication and real time persistence, Vercel for one click deployment with generated configs and environment variables, and Git for repository initialization, commits and pull request creation that returns a link. SourceVS Code Marketplace listing and gocodeo.com; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Workflow Orchestration | Partial |
|
Workflows chain several steps across connected tools. In GoCodeo's worked example, one flow fetches a PRD from Notion, generates the application code, creates a GitHub repository and deploys through Vercel. The product is split into modular BUILD, ASK, TEST and MCP flows, and the agent selects tools by intent. Enhanced prompts expand a request into functional and technical specifications before generation, and build history lets the developer switch between sessions. GoCodeo names no multi agent architecture, agent coordination or handoff in the product. A GoCodeo blog essay describes planner, builder, critic and deployer roles, but the product does not name them. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Knowledge Grounding & RAG | Partial |
|
Context comes from the active workspace. An @ symbol references specific files or code symbols from the open folder. The agent scans the repository to generate structured documentation, including project structure and dependencies. PDFs and Word documents can be uploaded and queried, and images can be attached for visual context and turned into code. An @Web search powered by Perplexity fetches live API references and documentation with sources, and suggestions draw on imported libraries and framework configuration. GoCodeo names no codebase index, semantic retrieval or persistent project understanding across sessions. Sourcemarketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-09-29 |
|
| Human Oversight & Guardrails | Partial |
|
The developer directs each step by prompt, reviews generated files in the editor and changes them with targeted section edits. Enhanced prompts show a functional and technical specification before generation, so the plan is visible first. Build history allows switching back to a previous build session, and pull request creation puts changes into normal human review. GoCodeo names no approval gate for each action, no permission scoping on agent actions and no configurable runtime guardrail, and deployment to Vercel is a one click action. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Security, Identity & Governance | Not documented |
|
Customer code is never stored or used for any purpose, including training, GoCodeo says, and it collects only anonymized telemetry and usage data. All generated code belongs to the customer, including for commercial use. For data handling, GoCodeo links to the privacy policies of OpenAI, Anthropic and Google, its model providers. It names no security certification or attestation, SSO, SAML, RBAC, audit logging or admin governance surface, and it publishes no DPA, security page or trust center. The site carries Terms of Use but no privacy policy. Sourcegocodeo.com/pricing, gocodeo.com footer and the VS Code Marketplace listingread 2026-08-30 |
|
| Observability & Auditability | Partial |
|
Tool calls show inline in chat. GoCodeo claims visibility into which tool was called and its result, and describes arguments, raw output and agent reasoning. It also says inline tool arguments and step by step execution logs are still upcoming, not shipped. Build History records past build sessions for review, test failures are explained in the editor, and generated code lands in the workspace as files and diffs the developer can inspect. GoCodeo names no retained execution trace, audit log or analytics surface outside the chat session. Sourcemarketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-09-29 |
|
| Memory & State Persistence | Partial |
|
Build History Management keeps multiple build sessions organized and switchable, so past iterations can be revisited and referenced. Chat history persists across multiple conversations, with session switching. Generated projects and their files persist in the local workspace and the customer's own Git repository. GoCodeo names no learned memory that builds up across sessions and no checkpoint or rollback module. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Deployment & Data Residency | Partial |
|
The VS Code extension runs locally, with a backend GoCodeo hosts for model calls. Generated code is written into the local workspace and the customer's own Git repository, and it belongs entirely to the customer, including for commercial use. Backends are provisioned into the customer's own Supabase project, and applications deploy to the customer's own Vercel account with generated configs and environment variables. GoCodeo names no self hosted, on premises, VPC or region selectable deployment of GoCodeo itself. It says it does not store code and collects only anonymized telemetry. Sourcegocodeo.com/pricing and the VS Code Marketplace listingread 2026-08-30 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
Framework starter templates cover 25+ frontend and backend stacks, including React, Next.js, Django, Flask, Express.js and Spring Boot, and generate optimized scaffolds with essential files and configurations. Generated test cases can be customized and saved for reuse. A Browse Community MCP Servers option opens a catalog of prebuilt tool integrations built by other developers. GoCodeo names no library of prebuilt agents, agent marketplace or shareable agent pack, and its modules are built into the product, not installed separately. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Triggers & Channel Coverage | Not documented |
|
Runs start with a manual action by the developer inside Visual Studio Code. Ctrl+Shift+K or Cmd+Shift+K opens BUILD, Ctrl+Shift+L opens ASK, Ctrl+Shift+T gives terminal command suggestions, Ctrl+Shift+D debugs a failed terminal command and Ctrl+Shift+E edits a selected section. A GoCodeo-TEST button above a class or function generates tests, with a Run Tests button beside it. GoCodeo names no event trigger such as a pull request, commit, CI failure or webhook, no scheduled run and no inbound channel such as chat, ticketing or email. There is no IntelliJ or JetBrains distribution. SourceVS Code Marketplace listing and gocodeo.com; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Model Flexibility & Routing | Full |
|
A Model Selector lets the developer choose the model behind the agent across four providers, OpenAI, Claude, Gemini and DeepSeek. Pricing splits them into premium and basic request pools and meters Claude 4 Sonnet separately, at ten requests per day on Pro. GoCodeo advertises bring your own API keys with no markup, but also says no API keys are required because models ship ready to use. Named models are Claude 4 Sonnet, Claude Sonnet 3.7 and 3.5, Gemini 2.5 Pro, GPT-4.1, GPT-4o, o3-mini and DeepSeek-R1. Sourcegocodeo.com/pricing and the VS Code Marketplace listing; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| APIs, SDKs & MCP Extensibility | Partial |
|
Customers extend the agent by defining their own tools, including internal APIs, in an mcp.json configuration that names the tool, endpoint, arguments and tokens. Tools can be added globally or for each project. Community built MCP servers can be browsed and installed from inside the product. The extension itself is distributed through the VS Code Marketplace. GoCodeo names no public REST API, SDK or MCP server that exposes it to third party agents or applications. It refers to a unified MCP client and server architecture but does not say what that exposes. Sourcemarketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-09-29 |
|
| Testing, Debugging & Optimization | Full |
|
TEST is a dedicated module with its own panel and one click Generate Tests and Run Tests controls. It generates unit tests covering positive, negative, edge and error cases from a class or function. The framework, code style and number of tests per run are all configurable, and generated cases can be saved and reused in an existing suite. There is a regenerate control, and a feedback loop refines future generation. AI driven explanations cover failures, and a Fix Test/Code control repairs them. Test execution covers Python and JavaScript. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
| Browser & Computer Use | Not documented |
|
The agent works entirely through programmatic surfaces. These are the VS Code extension API, a terminal helper and terminal debugger bound to keyboard shortcuts, file system writes, Git operations, Supabase and Vercel APIs for backend setup and deployment, MCP tool calls defined in mcp.json, and an @Web search powered by Perplexity. GoCodeo names no browser control, screenshot capture, visual verification or way to operate software that lacks a programmatic interface. SourceVS Code Marketplace listing and gocodeo.com/features; marketplace.visualstudio.com/items?itemname=gocodeo.gocodeoread 2026-08-30 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented 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
Pricing
Hobby free · Starter $9/mo · Pro $19/mo
Per user monthly subscription with a monthly quota of premium model requests and unlimited basic model requests; overage handled by a paid add on. Free tier with limited requests; enterprise custom.
Included quota
GoCodeo meters requests rather than seats. Hobby is free, with 50 premium model requests a month, 1,000 autocompletions, 100 basic model requests, 5 deployments and MCP integration. Starter costs $9 a month or $85 a year, with 200 premium model requests, 4,000 autocompletions, 400 basic model requests and 50 deployments. Pro costs $19 a month or $182 a year, with 500 premium model requests, 10 Claude 4 Sonnet requests a day, unlimited autocompletions and basic requests, and 100 deployments a month. Annual billing saves about 20 percent, and a $10 add on buys 250 more premium requests within a month on any paid plan.
What is public
Three tiers are published with monthly and annual prices and their quotas: Hobby free, Starter $9 a month or $85 a year, and Pro $19 a month or $182 a year. Per tier limits for premium requests, autocompletions, basic requests and deployments are listed, as is the $10 add on. No Business or Enterprise price is published, and the extension's marketplace listing describes a bring your own key model that the pricing page does not mention.
Billing mechanics
Per user monthly subscription. Each paid tier includes a monthly premium model request quota (500 on Pro) plus unlimited basic model requests; a $10 add on buys 250 more premium requests for the month. Free tier with limited requests.
Cost watchouts
Premium model requests (GPT-4o, Claude Sonnet, DeepSeek, o3-mini) are capped monthly (500 on Pro); heavy use requires a $10 add on for 250 more, so cost can climb with intensive premium model usage. Basic model requests are unlimited.
Variable cost rationale
The monthly subscription is predictable, but premium model requests are capped (500 on Pro) and intensive use of the strongest models pushes a user to $10 add on packs of 250 requests, so cost scales with how heavily premium models are used. Unlimited basic model requests bound the floor. Exposure is moderate: a predictable base with usage sensitive premium request overage.
Additional watchouts
No formal security certification such as SOC 2 is published. Premium model requests are capped monthly and heavy users will buy add ons. The product is young and from a small team, and its autonomy suits greenfield app building more than large legacy codebases. Cloud backed, so review the privacy stance for sensitive code even though the vendor states it does not store or train on user code.
Overage / add-ons
Premium model requests are capped per month (500 on Pro); exceeding the cap requires a $10 add on that grants 250 additional premium model requests for that month. Basic model requests are unlimited and not metered.
Sales call required
No, self serve available
Free / trial
Hobby plan, free: 50 premium model requests per month, 1,000 autocompletions, 100 basic model requests, 5 deployments and MCP integration.
Lowest paid plan
Starter $9/month, or $85/year
Commercial notes
Priced as an affordable, self serve, full lifecycle coding agent bundling build, test and deploy. Value is highest for individual developers and small teams doing greenfield full stack work who want autonomous scaffolding, testing, and one click Vercel and Supabase deployment. Multi model access is included rather than requiring a separate model subscription.
Key ambiguities
A first-party contradiction is unresolved and should be checked before quoting price. The VS Code Marketplace listing states GoCodeo is now completely free because the customer brings their own API keys, with no per-token charge, commission or markup, while the pricing page sells metered tiers and asks 'Why isn't GoCodeo completely free?'. The same marketplace page also states no API keys are required because models ship ready to use, contradicting its own BYOK claim. Separately, the pricing page lists only Hobby, Starter and Pro, but its FAQ refers to exceeding limits on 'any paid plan (Pro or Business)', so a Business tier may exist unlisted; no Enterprise pricing is published. Which premium models count is stated inconsistently, with the FAQ naming GPT-4o, Claude Sonnet 3.5, DeepSeek-R1 and o3-mini while the Pro tier separately meters Claude 4 Sonnet.
Cancellation / refund
Monthly self serve plans with a free tier. Add on premium request packs are one time purchases for the month. Specific cancellation and refund terms are not detailed.
Support SLA / resale
Self serve support on paid tiers with an enterprise option for larger teams. The vendor states it does not store or use code for training and collects only anonymized telemetry. No reseller or white label program surfaced.
Missing data
Business tier pricing, if the tier exists, is not published; Enterprise pricing and seat minimums are not documented.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose AI agents build, host and manage…
- AgentUI — Managed AI app builder for operations teams that generates and hosts…
- Aider — Open source, model agnostic terminal coding agent that edits your…
- Amp — Amp is a coding agent from Amp Frontier Corporation that runs each…
- Anthropic Claude Code — Anthropic's agentic coding system across terminal, desktop, IDE, web…
Alternatives to GoCodeo
The closest documented capability profiles to GoCodeo among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Codebuff9.0 / 14Adds documented Triggers & Channel Coverage and Browser & Computer Use
- Ideavo9.0 / 14Adds documented Security, Identity & Governance and Triggers & Channel Coverage, among others
- iGent5.0 / 14Adds documented Security, Identity & Governance
- OnSpace6.0 / 14Adds documented Security, Identity & Governance and Triggers & Channel Coverage
- Vibecode10.0 / 14Adds documented Triggers & Channel Coverage and Browser & Computer Use
- Aider9.5 / 14Adds documented Triggers & Channel Coverage
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded