imp

# Imp v0.17.0

Imp is a framework for typed, optimizable language-model programs on the BEAM.
This release sends a chat message's participant `name` (the OpenAI chat
format's optional `name`) to the provider, where Imp used to drop it, and lets
a program name the messages of its history turns and its request. A message
without a name is sent exactly as before, so no caller that does not name
messages sees any change. The new input keys are public API a caller can rely
on, so under Imp's versioning rule this is a minor release.

## Install

```elixir
{:imp, "~> 0.17.0"}
```

`~> 0.17.0` takes later 0.17 patch releases and nothing newer. The two-part
form, `~> 0.17`, means `>= 0.17.0 and < 1.0.0` in Mix, so `mix deps.update
imp` under it takes every later 0.x release, including minor releases that
change what callers receive.

Every dependency comes from Hex. Use a path dependency only while developing
against a local checkout.

ExMCP and erlexec are declared `runtime: false`, so an OTP release that uses
`Imp.ACP` or `Imp.MCP` must list `applications: [ex_mcp: :load, erlexec: :load]`
in its release definition; see [releases that use MCP or
ACP](docs/production.md#releases-that-use-mcp-or-acp).
Ordinary Imp startup starts no protocol endpoint.

## Changes and upgrading from 0.16.1

Change the Imp dependency to `{:imp, "~> 0.17.0"}`, run `mix deps.update imp`,
and commit `mix.lock`. No code or data migration is needed. A program that
already puts `name` on a message it gives `Imp.LM`, or `:user_name` in its
inputs, now sends it.

- **Added: participant names reach the provider.** `Imp.Clients.ReqLLM`
  copies a message's `name` (atom or string key) onto the message it sends,
  for generated, streamed and batched requests. Before, `Imp.LM.generate(lm,
  [%{role: "user", content: "hi", name: "alice"}])` went out with no `name`.
  A provider whose format has no such field, such as Anthropic's Messages
  API, does not receive it.
- **Added: naming messages through the adapters.** With `Imp.Adapter.Chat`,
  and the JSON, XML and PlanFirst adapters built on it, a history turn names
  its user message with `:user_name` and its assistant message with
  `:assistant_name`, beside its fields, and the request inputs name the final
  user message with `:user_name`:

  ```elixir
  history = Imp.History.new([%{question: "Who won?", answer: "Blue.", user_name: "alice"}])
  Imp.call(predict, %{question: "Rematch?", history: history, user_name: "bob"})
  ```

  The keys are never rendered as fields, and tool results and notes are
  never named. A signature that declares a field called `user_name` or
  `assistant_name` keeps it as an ordinary field. `Imp.Predict` does not warn
  about `:user_name` as an extra input under those adapters. ReActV2 names
  the turns of a history it is given, but not its request message.
- **Not every model sees the name.** Many upstreams accept `name` and drop
  it, among them Google's, DeepSeek's, Kimi's, Mistral's and Grok's, and
  open-weight models served through hosts such as DeepInfra. When the model
  must know who is speaking, say so in the content too.

Rollback means restoring `{:imp, "~> 0.16.0"}` and `0.16.1` in the lock;
names are then dropped again and `:user_name` in the inputs warns as an
extra input. See the [changelog](CHANGELOG.md) for every change.

## Known limits

- The usage map (`:usage` on the `:model_response` event, `Imp.Usage`,
  `Imp.Prediction.get_lm_usage/1`) is ReqLLM's, with `input_includes_cached`
  added. Its `:cost` and `:total_cost` are ReqLLM's catalog estimate, not a
  charge, for streamed and non-streamed calls alike, and OpenRouter's charge
  is its `"cost"`.
  Read money from `cost` and `estimated_cost`.
- An error the provider sends inside a stream has `status` `nil` and
  `retryable: true`, because ReqLLM's stream decoder keeps only its message;
  the same error in a non-streamed response may carry a status that says not
  to retry.
- A flat two-element name-first list that is not itself an element of a
  list (at the top level, in a tuple, or under a key that is not a credential
  name), such as `["password", "hunter2"]`, is not redacted, where 0.5.0
  redacted it. Such a list is read as a key and value only as an element of a
  list, so that lists of names such as `with_inputs([:api_key, :question])`
  survive saving.
- The GEPA, SIMBA, MIPROv2, InferRules and random-search checkpoints redact
  their failure reasons but otherwise hold the resume state as it is: GEPA's
  candidates, evaluation cache, proposed instructions and reflection data, and
  the others' instructions, demos and scores. Fast-Slow and Playbook
  checkpoints likewise hold their prompts, rollouts and playbooks. Treat
  checkpoint files as sensitive.
- A ReActV2 step whose prose quotes its own field names (a JSON object with
  `next_thought` or `tool_calls` keys, a `[[ ## field ## ]]` line, a
  `<next_thought>` tag) is not read as prose and goes to the JSON fallback,
  which usually costs one more call. If the model repeats the same reply, the
  step ends `:incomplete`. Beside native tool calls, such a thought becomes
  `nil`; the calls still run.
- Tool images reach the model as user messages after the turn's tool
  responses, labelled with the call, not nested in the tool result, even for
  providers that accept images in tool results. Only MCP text and image
  blocks are converted; other block types stay text, image URLs in text are
  never fetched, and the selected model must accept image input of that
  format and size.
- ATIF's `extra.seen_by_model` and the read-as `content` come from matching
  a tool message in the next captured request. A result that request does
  not carry, including one that ended the run, keeps the tool's own output
  and is marked unseen. `agent.tool_definitions` comes only from `:tools_sent`
  events, and `final_metrics` totals are left out when any response lacks
  that figure.
- A streamed completion that the provider's content filter stops still ends
  as a completion: the stream's `finish_reason` is `:content_filter` and the
  filter's text has already arrived as chunks. Only non-streamed requests
  turn it into an `Imp.LMError`.
- When a streamed ReqLLM request fails, ReqLLM's own warning log prints the
  error with its response headers. Imp strips them from the error it returns,
  but cannot change that log line.
- In command output, a `Bearer` token of 12 to 15 characters, or one with no
  digit, is shown when more text or a newline follows it. 0.5.0's command
  pipeline hid any such token of 12 or more characters.
- A GEPA checkpoint taken between preparing and starting a full validation
  resumes by rejecting that candidate, as before.

The [CHANGELOG](CHANGELOG.md) records every user-visible change in this
release. Generated module documentation is the complete API reference. Start
with `Imp`, `Imp.Signature`, `Imp.Module`, `Imp.Evaluate`, `Imp.Optimizer`,
`Imp.ACP`, `Imp.MCP`, and `Imp.Telemetry`.