Skip to content

Open source AI coding agent. Designed for large projects and real world tasks.

ollama quickstart, minor fixes for provider/env var docs

Dane Schneider9a01fea8cbed
diff --git a/docs/docs/environment-variables.md b/docs/docs/environment-variables.md
index 00dc1163..2fd07f95 100644
--- a/docs/docs/environment-variables.md
+++ b/docs/docs/environment-variables.md
@@ -41,6 +41,7 @@ VERTEXAI_LOCATION= # Your Google Vertex AI location
AZURE_OPENAI_API_KEY= # Your Azure OpenAI API key
AZURE_API_BASE= # Your Azure OpenAI API base URL
AZURE_API_VERSION= # Your Azure OpenAI API version
+AZURE_DEPLOYMENTS_MAP= # Your Azure OpenAI deployments map—a JSON object mapping model names to deployment names (only needed if deployment names are different from model names)
# DeepSeek
DEEPSEEK_API_KEY= # Your DeepSeek API key
@@ -49,9 +50,12 @@ DEEPSEEK_API_KEY= # Your DeepSeek API key
PERPLEXITY_API_KEY= # Your Perplexity API key
# Amazon Bedrock
+PLANDEX_AWS_PROFILE= # Name of AWS profile in ~/.aws/credentials to use for AWS Bedrock. If not set, the credentials file won't be used.
AWS_ACCESS_KEY_ID= # Your AWS access key ID
AWS_SECRET_ACCESS_KEY= # Your AWS secret access key
AWS_REGION= # Your AWS region
+AWS_SESSION_TOKEN= # Your AWS session token
+AWS_INFERENCE_PROFILE_ARN= # Your AWS inference profile ARN
```
### Upgrades
diff --git a/docs/docs/models/model-providers.md b/docs/docs/models/model-providers.md
index b5f81b08..cbef2bae 100644
--- a/docs/docs/models/model-providers.md
+++ b/docs/docs/models/model-providers.md
@@ -100,7 +100,17 @@ export AZURE_DEPLOYMENTS_MAP='{"gpt-4.1": "gpt-4.1-deployment-name"}' # optional
### AWS Bedrock
-You can optionally use AWS Bedrock for Anthropic models. AWS Bedrock uses standard AWS authentication via environment variables.
+You can optionally use AWS Bedrock for Anthropic models.
+
+If you have an AWS credentials file at `~/.aws/credentials`, you can use that to authenticate by setting the `PLANDEX_AWS_PROFILE` environment variable:
+
+```bash
+export PLANDEX_AWS_PROFILE=... # set the name of the profile in ~/.aws/credentials to use
+```
+
+Note that the credentials file will _only_ be read if `PLANDEX_AWS_PROFILE` is set.
+
+You can also use environment variables for AWS authentication:
```bash
export AWS_ACCESS_KEY_ID=... # set your AWS access key ID
diff --git a/docs/docs/models/models-overview.md b/docs/docs/models/models-overview.md
index 3c465caa..da74e56c 100644
--- a/docs/docs/models/models-overview.md
+++ b/docs/docs/models/models-overview.md
@@ -99,13 +99,4 @@ While you can use Plandex with many different providers and models as described
Plandex supports local models via [Ollama](https://ollama.com/). For more details, see the [Ollama Quickstart](./ollama.md).
-### Local Models Disclaimer
-
-While local models are supported via Ollama, small models that can be run locally often aren't strong enough to produce usable results for the [heavy-lifting roles](./roles.md) like `planner`, `architect`, `coder`, and `builder`. The prompts for these roles require strong instruction following that can be hard to achieve with small models.
-
-The strongest open source models _are_ capable enough for decent results, but these models are quite large for running locally without a very powerful system. This isn't meant to discourage experimentation with local models, but to set expectations for what is realistically achievable.
-
-To help bridge the gap as local models continue to improve their capabilities, a built-in `ollama-adaptive` model pack is available. This model pack uses local Ollama models for less demanding roles, plus larger remote models for heavy-lifting.
-
-
diff --git a/docs/docs/models/ollama.md b/docs/docs/models/ollama.md
index 182ab24a..4b6f51f4 100644
--- a/docs/docs/models/ollama.md
+++ b/docs/docs/models/ollama.md
@@ -5,4 +5,81 @@ sidebar_label: Ollama Quickstart
# Ollama Quickstart
-Plandex works with [Ollama](https://ollama.com/) models.
\ No newline at end of file
+Plandex works with [Ollama](https://ollama.com/) models. To use them, you need to [self-host Plandex.](../hosting/self-hosting/local-mode-quickstart.md) Ollama isn't supported with Plandex Cloud.
+
+## Disclaimer
+
+While local models are supported via Ollama, small models that can be run locally often aren't strong enough to produce usable results for the [heavy-lifting roles](./roles.md) like `planner`, `architect`, `coder`, and `builder`. The prompts for these roles require strong instruction following that can be hard to achieve with small models.
+
+The strongest open source models _are_ capable enough for decent results, but these models are quite large for running locally without a very powerful system. This isn't meant to discourage experimentation with local models, but to set expectations for what is achievable.
+
+To help bridge the gap as local models continue to improve their capabilities, a built-in `ollama-adaptive` model pack is available. This model pack uses local Ollama models for less demanding roles, plus larger remote models for heavy-lifting. There's also a built-in `ollama-experimental` model pack that uses local models for all roles—this is recommended for testing and benchmarking, but not for getting real work done.
+
+## Install and run Ollama
+
+[Download and install ollama](https://ollama.com/download) for your platform.
+
+Then make sure the ollama server is running:
+
+```bash
+ollama serve
+```
+
+## Pull Ollama models
+
+Pull the models you want to use. For the built-in `ollama-adaptive` and `ollama-experimental` model packs, pull the following models:
+
+```bash
+ollama pull qwen3:32b
+ollama pull qwen3:8b
+ollama pull qwen3:14b
+ollama pull devstral:24b
+```
+
+## Use Ollama in Plandex
+
+### Built-in model packs
+
+To use one of the built-in Ollama model packs in Plandex, decide whether you want to use `ollama-experimental`, which uses local models for all roles, but may struggle in practice, or `ollama-adaptive`, which uses local models for less demanding roles, plus the default Plandex models for heavy-lifting.
+
+```bash
+\set-model ollama-experimental # REPL
+plandex set-model ollama-experimental # CLI
+```
+
+Or:
+
+```bash
+\set-model ollama-adaptive # REPL
+plandex set-model ollama-adaptive # CLI
+```
+
+### Custom models and model packs
+
+You can also setup [custom models and model packs](./custom-models.md) for use with Ollama.
+
+When configuring a custom model, be sure you add the `ollama` provider to the `providers` array with the `modelName` set to the name of the model you want to use, exactly as it appears in the [Ollama model list](https://ollama.com/models), prefixed with `ollama_chat/`. For example, to use the `qwen3:32b` model, you would add the following to the `providers` array:
+
+```json
+"providers": [
+ {
+ "provider": "ollama",
+ "modelName": "ollama_chat/qwen3:32b"
+ }
+]
+```
+
+When configuring a custom model pack to use Ollama, set the top-level `localProvider` key to `ollama`. For example:
+
+```json
+{
+ ...
+ "modelPacks": [
+ {
+ "localProvider": "ollama",
+ ...
+ }
+ ]
+}
+```
+