Contentful Assistant

Slack
Contentful
Vercel

A Slack-based Contentful assistant that searches, drafts, and edits entries, with approval before publishing content.

Slack
Contentful
Vercel

agent/agent.tstypescript

import { defineAgent } from "eve";

export default defineAgent({
  limits: {
    maxInputTokensPerSession: 5_000_000,
    maxOutputTokensPerSession: 100_000,
  },
  model: "openai/gpt-6.1-sol-fast",
});

eve Contentful Agent Template

Slack-based Contentful assistant built on eve, with Vercel Drives providing a searchable copy of your content. Draft and edit entries directly in Slack, then review publication requests before the content goes live. The agent also handles asset uploads and can present results as tables or charts.

Getting started

Vercel's project setup flow will create a repository in your Git account and a Vercel project. You’ll create and install a Slack app via Vercel Connect, configure a private Vercel Blob store to save users’ writing preferences, and be prompted to add the following environment variables:

Variable	Value
`CONTENTFUL_MANAGEMENT_TOKEN`	Your Contentful Management API token, with access to your configured spaces
`CONTENTFUL_SPACE_IDS`	Your Contentful space ID, or multiple IDs separated by commas

Once you've deployed the agent, add it to a Slack channel and @mention it to start your first conversation.

Tech stack

Layer	Technology
Agent framework	eve
Content management	Contentful
Slack integration	Vercel Connect
Content search and sync	Vercel Sandbox
Per-user writing preferences	Vercel Blob

Publishing

The agent saves edits without publishing entries. By default, publication requires approval in Slack, where a card lists the entries and assets that will go live. Any authenticated user can approve or cancel the publication request.

You can exempt supporting content types, such as authors or callouts, from approval. New asset uploads are processed and published when added to content. See publication settings for these options and their effect on linked content.

User preferences

The agent remembers each user's writing preferences, such as tone or language, and stores them separately in private Vercel Blob storage. Preferences saved during local development are temporary.

Local development

Use Node.js 24 with pnpm for local development. Install the Vercel CLI, then clone your repository and run:

Choose your deployed project when vercel link prompts you. vercel env pull .env.local saves your project's development environment variables to .env.local.

pnpm dev opens eve's interactive chat in your terminal. Type a message such as "List the content types in my space" to talk to the agent, which reads and edits content in your configured Contentful space.

Local sessions use just-bash without access to the content stored on Vercel Drives, so searches query Contentful directly. Test searches using Drives and publication approval cards through the agent in Slack.

Before deploying, check your changes and confirm the agent builds:

Use pnpm fix to apply automatic formatting and lint fixes. When the checks pass, deploy with:

Customization

Customize	Where
Your team's writing guidelines	agent/instructions.md
Which content types need approval	`CONTENTFUL_COMPONENT_TYPES`
Website URL lookup	`CONTENTFUL_PUBLIC_ORIGIN` and `CONTENTFUL_PAGE_ROUTES`
Slack channels that receive replies without mentions	`SLACK_AUTO_REPLY_CHANNEL_IDS`
Model and session limits	agent/agent.ts

For more information on customizing the agent, see CUSTOMIZATION.md.