What does LangBot do? Do I need a ChatGPT account?

LangBot is an open-source IM bot platform that connects large models to WeChat, WeCom, Lark, and other chat channels. This page uses New API mode with an AI Code With key — you do not need to sign in to ChatGPT.

Why must Request URL include /v1?

LangBot's New API provider treats Request URL as an OpenAI-compatible base URL. The official docs require /v1. Omitting /v1 or using another vendor's default host causes 401 or scan failures.

The model test passed. Why does the pipeline still fail?

The Test button in Create Model only checks that one model record. The pipeline must select the same model and save. If the name does not match, the test can pass while live chat still returns Model Not Exist.

Where do I get the API key?

Create an API key in the AI Code With dashboard under Key Management. The key is shown only once, so save it, then paste it into LangBot's API Key field.

How do I confirm the setup works?

Save the model in the pipeline, then send a test message in WeChat or another connected channel. A normal reply confirms the connection; match the call by time and model in usage records.

LangBot × AI Code With

Use AI Code With in LangBot: third-party API / relay setup

This page shows how to connect a third-party API in LangBot so you can reply in WeChat and other chat platforms with a custom model.

Install LangBot first, then add New API in model configuration. Enter the key only in LangBot, never on this page.

I have LangBot: set it up Install LangBot

New API quick reference4 settings

Model provider: New API
Request URL: https://api.aicodewith.ai/v1
API Key: sk-xxxxxxxx
Model ID: deepseek-v4-flash

Quick setup

In LangBot Model Configuration, create a New API model: fill Request URL, API Key, and model name, then select it in the pipeline. About 3 minutes.

LangBot is an open-source IM bot platform. Open Model Configuration, create a model, choose New API, fill Request URL, API Key, and model name, then select the model in a pipeline.

1
Create a New API model
Open Model Configuration and create a model. Set the provider to New API. Set Request URL to https://api.aicodewith.ai/v1 (include /v1), paste your AI Code With API key, and paste a Model ID from this page into Model Name.
Quick copy
Request URL
https://api.aicodewith.ai/v1
Model provider
New API
Open the API key page
Copy a model name
The model name must match the AI Code With Model ID exactly. Copy an ID below, then paste it into Model Name.
Pick a family, then copy a Model ID into Model Name. Request URL stays the same for every model.
Selected
Provider
Claude
Model ID
claude-sonnet-4-6
The full catalog follows the AI Code With docs and pricing page. Example names in screenshots only show where to paste; use an ID from this page.
View the full model list
In Model Configuration, create a model. Choose New API, set Request URL to https://api.aicodewith.ai/v1, then paste the API key and a model name from this page.
Select the model in the pipeline
Open the pipeline, assign the model you just created to the built-in Agent, then save. Creating a model without selecting it in the pipeline does nothing.
Open the pipeline, assign the model you just created to the built-in Agent, then save. Creating a model without selecting it in the pipeline does nothing.
Send a message to verify
Send a test message in WeChat or another connected channel. A normal reply confirms the connection — then match the call by time and model in AI Code With usage records.
Send a test message in WeChat or another connected chat. A normal reply confirms the connection. Match the call by time and model in AI Code With usage records.

How does LangBot connect to AI Code With?

LangBot uses New API as an OpenAI-compatible provider. Connecting to AI Code With means changing Request URL, key, and model name, then selecting the model in a pipeline:

Create a New API model: Open Model Configuration and create a model. Set the provider to New API.
Fill Request URL: Set Request URL to https://api.aicodewith.ai/v1 (include /v1). This is the OpenAI-compatible address — do not use another relay domain.
Paste the model name: Paste a Model ID from this page into Model Name. LangBot does not fetch the AI Code With catalog.
Select it in the pipeline: Open the pipeline, assign the new model to the built-in Agent, and save. Creating a model is not enough.
Send a message: Send a test message in WeChat or another connected channel. A normal reply confirms the connection.

Flow: LangBot → New API (Request URL + API key + model name) → pipeline binding → AI Code With API. Everything is configured in the WebUI. No files to edit.

Frequently asked questions

New to LangBot?

LangBot is an open-source IM bot platform that connects large models to WeChat, WeCom, Lark, QQ, and other chat channels.

Official docs and GitHub

Finish deployment from the official docs, then open Model Configuration in the WebUI. This site does not host the install package.

Open the LangBot deploy docs

Full install guidance follows the LangBot official site. This page only covers connecting AI Code With.

Common error troubleshooting

The most frequent problems fall into 4 categories.

401 / wrong Request URL

Symptom: Create, scan, or test returns 401, Invalid Authentication, or an invalid API key.

Cause: Request URL was set to another vendor default, /v1 was omitted, or the wrong key was pasted.

Set Request URL to https://api.aicodewith.ai/v1 only — include /v1
Confirm the key is a valid AI Code With API key, not a LangBot lbk_ key
Re-paste the key with no extra spaces

Model Not Exist / name mismatch

Symptom: The model test in configuration may pass, but the pipeline or bot returns Model Not Exist or a request-parameter error.

Cause: The model name does not match the gateway ID, or the pipeline still points at an old model.

Copy the Model ID from this page instead of typing a screenshot example
Open the pipeline and confirm it selects the model you just created
Confirm the key can access that model

Pipeline did not bind the new model

Symptom: Logs show LLM model not found or not configured, or No LLM model configured for local-agent runner.

Cause: The model was created but not assigned to the built-in Agent, or the pipeline was not saved.

Open Pipeline → Configuration → Model and select the new model
Click Save, then test again
Do not stop at the Test button in the create-model dialog

Model test passes but chat has no reply

Symptom: The create-model test succeeds, but pipeline debug or WeChat chat does not reply normally.

Cause: The pipeline was not saved, the bot is not bound to that pipeline, or the live request still hits the wrong model.

Confirm the pipeline is saved and the bot uses that pipeline
Try another Model ID listed on this page
Check AI Code With usage records for a matching call

Need the full guide?

Every config parameter, the complete model list, and more troubleshooting live in the AI Code With docs.

Open the AI Code With docs