Pawmc: Private Assistant with Managed Context

A personal LLM assistant that runs on Cloudflare Workers and maintains a user-guided compounding knowledge base across conversations.

When starting a new conversation, choose which historical context you want it to have by choosing a category and topic. Additional context from new chats is compounded daily or on request.

It's simple and lightweight, and you own your data, pick your model, and control the prompts.

Key advantages

Compounding context that works

Pawmc asks for a little effort from a user, but the result is predictability of compounding context. Every time you start a chat, you must select a "category" and "topic" (existing or new) and that's exactly where the context will compound.

Over time, when you gather context about your fields of interest, you will have multiple context options to use as a background for your new chats. (e.g. "personal", "project-x", "hobby-y", "philosophy")

Simplicity

It's a great relief to use an LLM that doesn't have search-access and a system prompt of 200–1500 (!) lines. Pawmc's system prompt is one short sentence, focused on getting the point across.

Control over data

You deploy the software, you control the database. The database includes all your conversations and summaries. You can redact your summaries, you can redact your prompts, you can change the model.

In an ideal world, you would also control the LLM deployment, but as of mid-2026 most of us can't host a cutting-edge LLM on a private machine. It's the main tradeoff - prompts are still sent to the companies that host LLMs.

Cost

Deploying and hosting Pawmc on Cloudflare is way under their free tier limits, and you even get some free daily tokens with Workers AI.

Features

• Conversation Storage: Stores conversations by category and topic
• Context-Aware: Maintains AI-generated summaries for each category and topic
• Daily Summaries: Automatically updates summaries via cron daily
• API Key Auth: Secure access with Bearer token authentication
• Web Interface: Chat UI at the root URL

Setup

1. Create your own copy

Click "Use this template" on the GitHub repo to create your own repository.

2. Create a Cloudflare account

Go to dash.cloudflare.com/sign-up and create a free account. No credit card required.

3. Create an API token

1. Go to dash.cloudflare.com/profile/api-tokens
2. Click Create Token
3. Click Use template next to "Edit Cloudflare Workers"
4. Under Account Resources, select your account
5. Click Continue to summary, then Create Token
6. Copy the token — you'll need it in the next step

4. Add secrets to GitHub

In your new GitHub repository:

1. Go to Settings → Secrets and variables → Actions
2. Click New repository secret
3. Add these two secrets:

5. Deploy

1. In your GitHub repository click the Actions tab
2. Select the Deploy workflow from the left sidebar
3. Under three dots button on the right choose Run workflow option

Once deployed, you can put together a link to your worker, based on the email you used to register on Cloudflare, like in the example below.

replaceme@gmail.com => 
https://pawmc.replaceme.workers.dev

Open that URL, enter your API_KEY and start chatting.

Advanced: Use OpenRouter instead of Workers AI

By default, Pawmc uses Cloudflare Workers AI (free, no extra setup). However, Workers AI can be unreliable and is only ok for testing out the project. If you are planning to use it extensively, I recommend switching to OpenRouter, which is also supported.

Note that OpenRouter is pay-per-token — you'll need to add credit to your OpenRouter account (supports crypto as well). However, it's usually a symbolical spending due to project's simplicity and the budget-friendly model default.

To switch:

1. Create an account at openrouter.ai and generate an API key.
2. In your GitHub repository, go to Settings → Secrets and variables → Actions and add a new secret named OPENROUTER_API_KEY with your key.
3. In wrangler.jsonc, set AI_PROVIDER to "openrouter" and, if you want a different model, update AI_MODEL_OPENROUTER to any OpenRouter model slug.
4. Commit and push the change (or run the Deploy action) to redeploy.

To switch back, set AI_PROVIDER back to "workers-ai".

Getting updates

When the template repo is updated, go to your GitHub repo, click Sync fork → Update branch, and the action will redeploy automatically.

Free Tier Cloudflare Workers Limits