Harrison/quick start (#12491)

Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>

templates/docs/CONTRIBUTING.md

@@ -0,0 +1,43 @@
+# Contributing
+
+Thanks for taking the time to contribute a new template!
+We've tried to make this process as simple and painless as possible.
+If you need any help at all, please reach out!
+
+To contribute a new template, first fork this repository.
+Then clone that fork and pull it down locally.
+Set up an appropriate dev environment, and make sure you are in this `template` directory.
+
+Make sure you have `langchain-cli` installed.
+
+```shell
+pip install -U "langchain-cli[serve]"
+```
+
+You can then run the following command to create a new skeleton of a package.
+By convention, package names should use `-` delimeters (not `_`).
+
+```shell
+langchain hub new $PROJECT_NAME
+```
+
+You can then edit the contents of the package as you desire.
+Note that by default we expect the main chain to be exposed as `chain` in the `__init__.py` file of the package.
+You can change this (either the name or the location), but if you do so it is important to update the `tool.langchain`
+part of `pyproject.toml`.
+For example, if you update the main chain exposed to be called `agent_executor`, then that section should look like:
+
+```text
+[tool.langserve]
+export_module = "..."
+export_attr = "agent_executor"
+```
+
+Make sure to add any requirements of the package to `pyproject.toml` (and to remove any that are not used).
+
+Please update the `README.md` file to give some background on your package and how to set it up.
+
+If you want to change the license of your template for whatever, you may! Note that by default it is MIT licensed.
+
+If you want to test out your package at any point in time, you can spin up a LangServe instance directly from the package.
+See instructions [here](LAUNCHING_PACKAGE.md) on how to best do that.

templates/docs/INDEX.md

@@ -0,0 +1,13 @@
+# Templates
+
+A list of all template repos
+
+⭐Retrieval Augmented Generation Chatbot: Build a chatbot over your data. Uses OpenAI and Pinecone.
+
+⭐Extraction with OpenAI Functions: Do extraction of structured data from unstructured data. Uses OpenAI function calling.
+
+⭐Local Retrieval Augmented Generation: Build a chatbot over your data. Uses only local tooling: Ollama, GPT4all, Chroma.
+
+⭐OpenAI Functions Agent: Build a chatbot that can take actions. Uses OpenAI function calling and Tavily.
+
+⭐XML Agent: Build a chatbot that can take actions. Uses Anthropic and You.com.

templates/docs/LAUNCHING_PACKAGE.md

@@ -0,0 +1,41 @@
+# Launching LangServe from a Package
+
+You can also launch LangServe directly from a package, without having to pull it into a project.
+This can be useful when you are developing a package and want to test it quickly.
+The downside of this is that it gives you a little less control over how the LangServe APIs are configured,
+which is why for proper projects we recommend creating a full project.
+
+In order to do this, first change your working directory to the package itself.
+For example, if you are currently in this `templates` module, you can go into the `pirate-speak` package with:
+
+```shell
+cd pirate-speak
+```
+
+Inside this package there is a `pyproject.toml` file.
+This file contains a `tool.langchain` section that contains information on how this package should be used.
+For example, in `pirate-speak` we see:
+
+```text
+[tool.langserve]
+export_module = "pirate_speak.chain"
+export_attr = "chain"
+```
+
+This information can be used to launch a LangServe instance automatically.
+In order to do this, first make sure the CLI is installed:
+
+```shell
+pip install -U "langchain-cli[serve]"
+```
+
+You can then run:
+
+```shell
+langchain hub start
+```
+
+This will spin up endpoints, documentation, and playground for this chain.
+For example, you can access the playground at [http://127.0.0.1:8000/playground/](http://127.0.0.1:8000/playground/)
+
+![playground.png](playground.png)