I had a problem while setting up my chatbot so I got solution. Ths is a very common setup when you want Flowise (frontend / orchestration) to talk to Ollama (model server) inside Docker.

Let’s go step by step.

---

🧠 Concept

You have:

• Container 1: Flowise (runs on port 3000 by default)
• Container 2: Ollama (runs on port 11434 by default)

You want Flowise to send requests to Ollama from inside Docker, not just from your host machine.

---

🛠️ Solution: Put Them on the Same Docker Network

1. Create a User-Defined Bridge Network

This lets the containers talk to each other by name.

docker network create ai-network

---

2. Run Ollama in That Network

docker run -d \
  --name ollama \
  --network ai-network \
  -p 11434:11434 \
  ollama/ollama

This exposes Ollama on http://ollama:11434 inside the Docker network (and http://localhost:11434 from your host).

---

3. Run Flowise in That Network

docker run -d \
  --name flowise \
  --network ai-network \
  -p 3000:3000 \
  flowiseai/flowise

Now both containers share the same virtual network and can see each other by name.

---

4. Configure Flowise to Use Ollama

In Flowise UI:

• Go to Credentials → Create a new Ollama connection.
• Use the internal container name as the Base URL:

http://ollama:11434

(Do not use localhost or 127.0.0.1 — those refer to the Flowise container itself, not your Ollama container.)

---

5. Test the Connection

• In Flowise, drop an Ollama node in your workflow.
• Select the new credential with http://ollama:11434.
• Test by sending a prompt — you should get a response from the Ollama model.

---

🧩 Common Pitfalls

Problem	Fix
Flowise says “Fetch failed”	You are likely using localhost:11434. Use http://ollama:11434 instead.
Containers cannot talk	Check docker network inspect ai-network to see if both containers are listed.
Ollama not reachable from host	Make sure you used -p 11434:11434 so it’s exposed on your machine.
Firewall blocking	Open port 11434 if you are trying to connect from another computer.

---

16