Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 39 additions & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -130,6 +130,45 @@ deliberately over-sized request is rejected with the limit named in the error,
resolving in well under a second — so the `/<max>` shows up on the first turn
rather than after a multi-second `/v1/models` round-trip that returns nothing.

#### AWS Bedrock (via LiteLLM proxy)

minion works with any OpenAI-compatible endpoint, so you can use AWS Bedrock
models by running [LiteLLM](https://github.com/BerriAI/litellm) as a local
proxy:

```bash
pip install 'litellm[proxy]'
litellm --model bedrock/us.anthropic.claude-sonnet-4-20250514-v1:0 --port 4000
```

Then point minion at it:

```
MINION_BASE_URL=http://localhost:4000/v1
MINION_MODEL=bedrock/us.anthropic.claude-sonnet-4-20250514-v1:0
MINION_API_KEY=sk-noop
```

LiteLLM reads your standard AWS credentials (`~/.aws/credentials`,
`AWS_ACCESS_KEY_ID` / `AWS_SECRET_ACCESS_KEY`, or SSO/instance profiles).
Set `AWS_DEFAULT_REGION` if needed. Bedrock requires
[inference profiles](https://docs.aws.amazon.com/bedrock/latest/userguide/inference-profiles-support.html)
for on-demand throughput — use the `us.` or `global.` prefixed model IDs
(e.g. `us.anthropic.claude-opus-4-6-v1`) rather than bare model IDs.

A convenience wrapper that auto-starts the proxy:

```bash
#!/bin/bash
# save as ~/bin/minion-bedrock and chmod +x
export AWS_DEFAULT_REGION=us-east-1
if ! lsof -i :4000 &>/dev/null; then
litellm --model bedrock/us.anthropic.claude-opus-4-6-v1 --port 4000 &>/dev/null &
sleep 3
fi
exec minion "$@"
```

#### Built-in `together` source

If `TOGETHER_API_KEY` is set (in `~/.env` or your shell), minion auto-registers
Expand Down