This is the third tutorial in the Getting Started with Engaggit series. In the previous tutorial, you installed the app and finished onboarding. Now it's time for the one setup step Engaggit genuinely can't work without: connecting an AI provider.
Why You Need an AI Provider
Engaggit doesn't guess — it thinks. Every matched post gets read, scored, and turned into a starting point for a reply. That thinking is powered by AI, and it's what makes the whole product useful:
- AI scores every post for relevance (0–10). It reads the title, the body, and the comments, then tells you which conversations are genuine opportunities and which are noise.
- AI explains why a post matched. You get a sentiment label (positive, neutral, negative) and a match reason, so you understand what the AI saw before you spend a second of your time.
- AI generates comment drafts. When a post is a good fit, it writes a natural, helpful reply for you to review, edit, and post.
There is no "offline mode" that skips this. No AI provider means no relevance scores, no drafts, no Engaggit. That's why this is the first required step — but it's also the easiest, because you have 7 ways to do it, one of which is completely free.
Your 7 Options
Engaggit works with 7 AI providers. Don't panic about choosing — you can switch anytime in Settings with zero data loss. Here's the landscape at a glance:
| Provider | Cost | Setup Difficulty | Best For |
|---|---|---|---|
| Ollama (Local) | Free | Medium | Privacy, unlimited use, no API key needed |
| OpenAI | Paid (pay-per-token) | Easy | Best quality drafts |
| Gemini | Free tier / Paid | Easy | Google AI, good quality |
| DeepSeek | Free tier / Paid | Easy | Cost-effective cloud AI |
| Grok | Paid | Easy | X.ai's AI |
| OpenRouter | Free models available | Easy | One key for hundreds of models |
| Custom | Varies | Advanced | Your own AI endpoint |
A note about Ollama. You may remember seeing two Ollama entries in the onboarding overview — that's because Ollama comes in two flavors: local (runs on your computer, fully offline and unlimited) and cloud (Ollama's hosted models, no powerful hardware needed, with a free tier for
gemma4:31b-cloud). We cover both below. For everything else on this list, "free tier" means you can start without paying.
Getting an API Key (Step-by-Step)
Let's set up each option. The free ones first.
Option 1: Ollama (Free Local)
Ollama runs AI models directly on your computer. No account, no API key, no usage limits — the only cost is a bit of RAM.
- Go to ollama.ai and download the installer for your operating system.
- Install it like any other app (Windows
.exeor macOS.pkg). - Open a terminal and pull a model with one line:
(We recommend the localollama pull gemma4:26bgemma4:26bmodel — it's a great balance of quality and speed. You can pull any model you like instead.) - Leave Ollama running. It listens at
http://localhost:11434by default — that's the URL you'll use in Engaggit.
That's it. No signup, no key, no card. If you chose the Ollama cloud option instead, pick the gemma4:31b-cloud model — it gives you quality results without needing a powerful computer, and Ollama even offers a free tier for it, so you can still get started without paying. It follows the same "create an account and copy an API key" flow as the providers below.
Option 2: OpenAI
- Go to platform.openai.com and sign up or log in.
- Open the API Keys page.
- Click Create new secret key, give it a name, and create.
- Copy the key immediately — it starts with
sk-and is only shown once.
Option 3: Gemini (Google)
- Go to ai.google.dev.
- Get a key from Google AI Studio (the link is right there on the page).
- Copy the key.
Option 4: DeepSeek
- Go to platform.deepseek.com and sign up.
- Open the API Keys section and create a new key.
- Copy it.
Option 5: Grok
- Go to console.x.ai and sign in with your X account.
- Create a key under API Keys.
- Copy it.
Option 6: OpenRouter
- Go to openrouter.ai and create a free account.
- Open Keys and create a new one.
- Copy it. This single key unlocks hundreds of models — you'll pick the model inside Engaggit.
Option 7: Custom (Advanced)
If you run your own inference server (or a proprietary model) with an OpenAI-compatible endpoint, this option is for you. You'll enter your endpoint URL and any key it requires. It's powerful, but you should be comfortable with APIs before choosing it.
Security note: treat your API key like a password. Anyone with it can spend your money or use your quota. Paste it straight into Engaggit, don't share it, and if you think it's been exposed, regenerate it on the provider's site.
Entering Your API Key in Engaggit
Once you have a key (or Ollama running), connecting takes under a minute:
- Open Engaggit and go to Settings → AI Providers.
- Select your provider from the list.
- Paste your API key. (For local Ollama, skip the key and just confirm the URL is
http://localhost:11434.) - Choose a model from the dropdown or type by yourself — e.g.,
gpt-5.6-terra,gemini-3.5-flash-lite, or one of the Ollama models you pulled. - And Save
That's it — your AI provider is live, and Engaggit will use it for every scan from now on.
You can configure several providers at once — add as many as you like. Only one is active at a time; just pick which one Engaggit should use, and you can switch between them anytime in Settings → AI Providers. Handy if you want to run your Interests on one provider during the day and a cheaper one overnight.
Your API key is stored locally on your machine, not on Engaggit's servers.
Picking the Right Model
If you're not sure which model to pick, here's a cheat sheet based on what matters most to you:
| Your priority | Go with |
|---|---|
| Lowest cost | DeepSeek, or Gemini's / OpenRouter's free models |
| Best draft quality | OpenAI GPT-5.6 or Gemini 3.5 |
| Privacy, offline, unlimited | Ollama local |
| Flexibility (test many models) | OpenRouter |
Two things worth knowing:
- Bigger isn't always better to start. A small, cheap model like
gpt-5.6-terraorgemini-3.5-flash-liteis fast, affordable, and plenty good for relevance scoring. Upgrade only if you notice drafts lacking polish. - You can change anytime. Pick a reasonable default now, run a few scans, and swap models in Settings whenever you want. Nothing is lost when you switch — your Interests, keywords, and history stay put.
Troubleshooting Connection Errors
Most connection problems fall into three buckets:
| Error | What it usually means | Fix |
|---|---|---|
| "Invalid API key" | Wrong key, a typo, or the wrong provider selected | Double-check the copy/paste, verify the provider dropdown matches where the key came from, or regenerate the key |
| "Cannot connect to Ollama" | Ollama isn't running, or the URL is wrong | Make sure Ollama is open, and confirm the URL in Settings is http://localhost:11434 |
| "Rate limited" | You've hit your provider's request limit | Check your provider's quota, and reduce scan frequency in Settings if needed |
If you're still stuck after checking those, verify your internet connection is working — a blocked or offline network will produce connection errors on any provider.
What's Next
Your AI provider is connected. Now Engaggit is ready to think — but it still doesn't know what to look for. That's the job of an Interest: the configuration that tells Engaggit who you're trying to reach on Reddit and what you can offer them.
Next tutorial → Creating Your First Interest — the most important setup step, and where all your results come from. Coming soon
Frequently Asked Questions
1. Is there really a free option?
Yes. Ollama local is free forever — it runs on your own computer, needs no API key, and has no usage limits. Beyond that, Gemini, DeepSeek, and OpenRouter all offer free tiers you can start with. You genuinely don't need to spend money to try Engaggit.
2. Where is my API key stored?
Locally, on your machine. Engaggit is a desktop app with a zero-server architecture — your key, Interests, and data never leave your computer. It's not uploaded to any Engaggit server.
3. Does Engaggit send my Reddit posts to my AI provider?
Only the posts that match your search rules get sent to the AI provider you chose, for scoring and draft generation — that's the provider's job. If you want absolute privacy, choose Ollama local, which runs entirely on your machine, and nothing is sent anywhere at all.
4. Can I switch providers after setup?
Yes, anytime, from Settings → AI Providers. It takes seconds, costs nothing, and you lose no data. Many users A/B test providers — run the same Interest with two different providers and compare the match quality.
5. My connection works, but the drafts look generic. Why?
Two likely causes. First, the model you picked may be a small/cheap one — trying a higher-quality model can make an immediate difference. Second, draft quality depends heavily on how well you describe what you offer. The "What can I offer" field is your lever; we'll cover it in the next tutorial.
6. I have a key from another AI tool. Can I use it in Engaggit?
Yes — API keys are provider-specific, not tool-specific. If you already have an OpenAI, Gemini, DeepSeek, Grok, or OpenRouter key, just paste it into the matching provider in Engaggit.
Engaggit is a privacy-first, AI-powered Reddit lead generation app. It runs on your machine, keeps your data to yourself, and helps you never miss a relevant conversation again.