GitHub - TerseAI/Terse: Typescript Workflow builder, optimized for use in Coding Agents

Workflow building for the coding agent era.

The AI workflow platform for coding agents.
Connect your tools, generate a typed SDK from your workspace, and deploy serverless workflows that mix deterministic tool calls with agentic loops.

Quickstart · SDK · CLI · Sign up

Why Terse

AI agents are fast to build but make mistakes. Deterministic workflows are reliable but slow to set up. Terse lets you mix both in the same workflow: call tools directly when you need precision, hand off to an agent when you need judgment.

• Code-first. Workflows are TypeScript in your repo. Review them in PRs, track them in git history.
• Typed SDK from your workspace. terse generate generates strongly-typed handles for the channels, repos, projects, and tools you actually use, so the LLM (and your IDE) only sees what's in scope.
• Mix tools and agents. Call a tool deterministically when you know what you want. Hand off to an agent when you need judgment. Same workflow, same types.
• Scoped skills, no surprises. Tool access is bounded by the skills you declare. The agent can't reach outside that scope. No more catastrophic deletions caused by hallucinations.
• Secrets and OAuth, handled. Integration credentials live in Terse's secret manager (using Google Secret Manager) with automatic OAuth refresh.
• Serverless deploys. terse deploy ships your workflows to the Terse Cloud data plane (isolated Modal sandboxes). Every run is logged with a complete action trail.
• Run it where you want. Terse splits into a control plane (orchestrator, triggers, dashboard, history) and a data plane (your job code). Each runs on Terse Cloud or in your network, independently — see Deployment below.

Quickstart

Get a workflow live in under 5 minutes — full guide at docs.useterse.ai/quickstart.

With Code

With your coding agent (Claude Code, Cursor, Codex, and more)

npx terse-cli install         # one command: global CLI + agent skills + login

# in claude
/terse-create describe your workflow

npx terse-cli install
terse init my-project
cd my-project
code . # build some workflows
terse deploy

That's it. terse deploy ships to the Terse Cloud data plane — no infra to stand up.

Deployment

Terse has two planes you can place independently. Most users pick one of the named tiers below.

npx create-terse clones the repo, mints secrets, brings up Postgres in Docker, runs migrations, starts the app, and writes the env so terse-cli already points at your self-hosted control plane.

What a workflow looks like

A Terse workflow is a single TypeScript file. The job below watches a repo for new pull requests, posts a deterministic Slack message announcing the PR, then threads a Block Kit summary under it:

import { generateText, createJob } from "terse-sdk"
import { GithubPRTrigger, Repos, Skills, SlackChannel, Triggers, toolbox } from "../terse.generated"
// ^^ Generated based on your workspace

createJob({
    name: "Summarize PR and send slack message",
    triggers: [Triggers.github.onPROpened({ repo: Repos.TerseAI.Terse })],
    onTrigger: async (event: GithubPRTrigger) => {
        // Deterministic call — fixed channel, fixed message. No agent needed.
        const message = await toolbox.slack.sendMessage({
            channelId: SlackChannel.AllTerseInc.channelId,
            message: "New PR from " + event.sender.login + "!"
        })

        // Outputs are strongly typed
        const parentId = message.message_ts

        // Agentic: one prompt in, the model uses the GitHub + Slack tools granted via skills
        await generateText({
            prompt: `
            Summarize this PR ${event.formatForAgentRunner()}.
            Keep it short. Format the summary in Block Kit; include screenshots or diagrams from the PR as image blocks.
            Reply in a thread to the Slack message (thread parent ts: ${parentId}).
            `,
            skills: [
                // Fine tune what the agent has access to. Impossible to touch anything outside this scope
                Skills.github({ repos: [Repos.TerseAI.Terse] }),
                Skills.slack({ channel: SlackChannel.AllTerseInc })
            ]
        })
    }
})

The skills passed to generateText are the only surface the model can reach: it can post in one Slack channel and read this one repo, nothing else. The deterministic toolbox.slack.sendMessage call returns a typed result you can thread on directly, and generateText picks up the GitHub and Slack tools without further wiring.

Integrations

First-class, typed integrations for the tools you already use:

What's in this repo

This is the Terse monorepo — the platform code Terse runs in production, plus the public packages used to author workflows.

Documentation

• Docs: docs.useterse.ai
• Quickstart: docs.useterse.ai/quickstart
• SDK reference: docs.useterse.ai/reference/typescript-sdk
• CLI reference: docs.useterse.ai/reference/cli

Contributing

Issues and PRs welcome. If you're picking up something non-trivial, open an issue first so we can sanity-check the approach. See CONTRIBUTING.md for the full guide and SECURITY.md for the private disclosure process.

Working on the SDK, CLI, or frontend? Clone the repo and:

pnpm install
pnpm run dev             # watches terse-types, terse-sdk, terse-cli, frontend, backend
pnpm run install-global  # links your local terse-cli as `terse`

The frontend and CLI talk to the hosted Terse backend at api.useterse.ai by default, so most SDK/CLI/frontend work doesn't need a local backend. Override with TERSE_BACKEND_URL if you're pointing at a self-hosted instance.

Backend work needs Postgres, Redis, and the credentials backend integrations expect. See docs.useterse.ai/self-hosting for the full list.

Redis (required for the backend)

The backend uses Redis for pub/sub, Socket.IO fan-out, and rate limiting (the durable job queue runs on pg-boss in Postgres). It's a hard requirement: pnpm run dev exits with ECONNREFUSED errors if Redis isn't reachable. Start one locally before running the backend:

# Docker (recommended)
docker run -d --name terse-redis -p 6379:6379 redis:7

# or Homebrew (macOS)
brew install redis && brew services start redis

Set the connection string in backend/.env (this is the default, so it's only needed if your Redis runs elsewhere):

REDIS_URL=redis://localhost:6379

License

Terse is source-available under the Sustainable Use License v1.0. You can read it, fork it, modify it, and self-host it for non-commercial or internal business use. You can't redistribute it commercially or run it as a competing hosted product. The license text in LICENSE.md is authoritative.

Links

• Website: useterse.ai
• Docs: docs.useterse.ai
• LinkedIn: linkedin.com/company/terse-inc